The Reality of Tech Interviews in 2025

Interview processes are changing in a tech market that’s both cooling AND heating up at the same time. A deepdive with former Meta Staff Engineer/Engineering Manager Evan King and Stefan Mai

Hi – this is Gergely with the monthly, free issue of the Pragmatic Engineer Newsletter. In every issue, I cover challenges at Big Tech and startups through the lens of engineering managers and senior engineers. If you’ve been forwarded this email, you can subscribe here.

It’s been widely reported that the tech hiring market is much cooler than in 2020-2022; the number of software engineering job openings is down internationally in all major regions and the number of full-remote roles is in steady decline. Meanwhile, other metrics indicate that tech hiring is starting to recover – at least for senior engineers – as covered last month in the article, State of the startup and scaleup hiring markets, as seen by recruiters. It all adds up to a state of flux for candidates and employers to navigate through.

This article is an attempt to get clarity about how tech interviews are changing, by focusing on what the engineers who take interviews are seeing. For this, I turned to Evan King and Stefan Mai, cofounders of interview preparation startup, Hello Interview. Before starting it, Evan was a staff engineer at Meta for 4 years, and Stefan an engineering manager at Amazon for 6 years, and also a senior engineering manager at Meta. They’ve conducted hundreds of interviews, while Stefan has also been a hiring manager. Since launching their new business, they’ve helped thousands of engineers prepare for interviews, and have collected information on the pulse of the job market.

I reached out to them after reading their practical, fresh take on system design interviews, for candid takes on devs interviewing at startups and Big Tech in the current climate, especially compared to a few years ago. Today, we cover:

1. New reality of tech hiring. A rebounding market still well below its 2021-2022 peak.

2. Analyzing the tech hiring market. Artificial Intelligence (AI) and related sectors are hot, while frontend/backend/mobile are not. It’s tougher for new grads than experienced engineers.

3. Interview process changes. The formats of DSA and system design interviews remain the same, but are more demanding. Downleveling is more common, and team matching has quietly become another hurdle to clear.

4. Interview formats differ between startups and Big Tech. Startups embrace more practical interviews and AI tools, while Big Tech seems less flexible about changing approach.

5. Preparation strategies by experience. Advice for entry-level, mid-level, senior, staff+ tech professionals, and for EMs.

6. Silver linings. Big Tech hiring is up, there’s a boom in AI positions, and the playbook of interviews is public.

In the article below we’ll cover how demand for engineers in the AI field keeps being strong. For pointers on picking up engineering practices, see the deepdives AI Engineering in the real world and AI Engineering with Chip Huyen.

With that, it’s over to Evan and Stefan:

---

1. New reality of tech hiring

Three years ago, if you were a competent software engineer with 3+ years of experience, you likely had recruiters flooding your inbox with opportunities. Companies were fighting over engineering talent, throwing extraordinary compensation packages at candidates, and in some cases even looking past poor interview performance in order to secure hires faster. The 2020-2021 tech hiring frenzy was exceptional; a period many now look back on with a mixture of nostalgia and disbelief.

Fast forward to 2025, and the landscape has transformed dramatically. As co-founders of HelloInterview.com, we've had front row seats to these changes, observing tens of thousands of engineering interview journeys across companies of all sizes. In this deepdive, we aim to give you the unvarnished reality of tech interviewing in 2025, via real experiences of candidates navigating it today.

We’ve observed Big Tech’s hiring volumes are up roughly 40% year on year. This data comes from candidates currently working at late-stage companies, of whom the overwhelming majority use our platform to prepare for interviews they already have scheduled. This provides a reliable proxy for overall tech hiring trends, as candidates on our platform have immediate, concrete interview dates. An uptick in candidates getting more interviews suggests that the worst of the 2022-2023 tech winter has passed, and that there are more attractive openings worth preparing for. Still, we're operating in a fundamentally different market with new rules, expectations, and challenges.

The 40%-rebound figure is only part of the story. Yes, tech hiring is slowly making a comeback in aggregate terms, but it's a selective, strategic recovery that leaves some qualified engineers struggling to navigate processes which are now more demanding and less forgiving. Companies once desperate to fill seats are now being methodical and cautious, prioritizing precision in hiring decisions over speed and volume.

What we're witnessing isn't simply a market correction; it’s a subtle yet significant shift in evaluation standards. While the core interview structure at Big Tech remains largely unchanged, the bar has shifted approximately one standard deviation higher across the board, and performance that would have secured an offer in 2021 might not even clear the screening stage today.

2. Analyzing the tech hiring market

Here’s our take on the current job market.

Selective recovery

By the raw numbers, tech hiring appears on a solid upward trajectory. TrueUp.io's job trend tracking shows tech job postings have risen from a 2023 low of 163,000, to approximately 230,000 today; roughly a 41% increase.

Number of open tech jobs at tech startups, tech unicorns, and public tech companies. Source: TrueUp

The 42% increase in openings is consistent with what we've observed internally in HelloInterview usage metrics and mock interview volume, when we adjust data for candidates with interviews scheduled.

We are still well below the feverish heights of 2020-2022, though. Back then, open roles peaked at close to 500,000. The current recovery, while significant, has only restored us to around 46% of that peak.

Unlike previous tech hiring cycles when a rising tide lifted all boats, today's market is characterized by extreme selectivity. Companies have become far more picky about where they invest headcount, with major differences in opportunity based on specialization, experience level, and the prestige of ex-employers.

Domain specialization

Engineers in certain areas of specialization are seeing a lot of relevant openings, such as in:

• AI infrastructure

• Machine learning operations

• Generative AI application development

These areas of hiring are reminiscent of the 2021 peak; often with multiple offers, aggressive compensation, and expedited interview processes.

For example, a Bay Area staff engineer specializing in AI infrastructure at Google recently received a competing offer from Meta's AI infrastructure team which was above $1 million in total compensation. Previously at Google, numbers like this were typically reserved for senior staff positions. But getting a large pay bump when changing companies, while staying at the same level, is not an isolated incident; we're seeing similar-sized compensation packages for specialists in high-performance computing, ML systems design, and those specializing in responsible AI development.

Engineers in “core” domains see fewer opportunities. “Core domains” refers to frontend, backend services, mobile development, and similar areas. Later-stage startups that previously maintained multiple teams in these areas have consolidated with more empowered full-stack engineers. Focusing on full-stack leads to lower overall headcount, fewer openings, and more selective hiring processes. We see candidates with strong backgrounds in these areas often taking a long time to land a role, and when they do get an offer, the comp growth is rarely above what they currently earn. Note from Gergely: we previously saw how native mobile engineers face a tougher job market, and that becoming more full-stack is a sensible strategy for being more employable.

Senior engineers can still attract multiple offers, especially those with directly relevant experience for hiring companies. This could be deep domain expertise (e.g. working in the infrastructure domain when interviewing for infra teams, working in the finance domain when interviewing with FinTechs, etc), or it could be a deep technology expertise which matters to the employer. Meanwhile, engineers with less transferable skills face an uphill battle. Narrow skillsets often develop from working at the likes of Google or Meta, where people specialize narrowly in proprietary systems, tools, and technologies that don’t exist in the broader market.

Experience level divide

The current market is also starkly stratified by career stage, creating dramatically different realities for engineers depending on experience:

Junior engineers and new grads face the biggest challenge. We spoke with a job seeker based in India who graduated from IIT – the most prestigious computer science university in the country. They shared a meticulously-maintained spreadsheet of their job search:

• 6 months of searching

• 100 companies contacted; all known to hire from IIT

• 4 initial interviews

• Zero offers

Companies that once maintained robust university hiring programs have dramatically scaled them back, which is concerning because this could create an experience gap that impacts the industry for years to come, and could manifest in a “missing generation” of engineers. This could create an industry-wide shortage of early and mid-career talent; potentially stalling innovation, as fewer fresh perspectives enter the field and challenge established practices.

Mid-career engineers: more interview loops to an offer. By mid-career engineers we refer to professionals with around 3-4 years of experience at respected companies. Candidates with this background are generally securing interviews, but the number of interview loops they go through to get an offer has increased substantially.

For example, a high achieving, mid-level engineer in the Bay Area with 4 years of experience at Amazon went through eleven full interview loops at different scaleups and tech companies before receiving their first – and only – offer!

Senior and staff engineers with high-demand specializations: premium comp and multiple offers. Companies are willing to pay significantly above market rate for proven expertise in AI, infrastructure, and security. Such candidates often have the luxury of choosing between competing offers and negotiating aggressively.

One of Evan’s recent mentees is a principal SDE, based in the Bay Area, working in one of Microsoft's AI infrastructure groups. This Principal SDE received competing offers from NVIDIA, Snowflake, Meta, and other places – all within a single month!

Engineering managers face a tough market. Widespread organizational restructuring swept through tech in 2022-2023, eliminating entire management layers, and companies have been slow to restore these positions since. As a result, many qualified engineering leaders now compete for a significantly reduced pool of opportunities. The heightened competition has transformed hiring standards.

Technical abilities once overlooked for managers are now meticulously evaluated, and system design skills are also becoming non-negotiable. In the past, managers were often hired primarily on leadership capabilities, but today, they need to prove leadership, as well as being hands-on with technology, software engineering, and software design.

The priorities in leadership roles have also shifted dramatically. Many tech companies previously focused on big-organization skills to build alignment across large teams. Today, those companies seek senior leaders who can remain focused on execution and support the higher executive layer; they’re usually not looking for senior leaders who want to remain at the high-level, interested in only steering the ship.

This transition to engineering leaders being expected to be hands-on feeds into longstanding debates about the distinction between engineering managers and technical leads. After all, today’s engineering leaders look awfully similar to yesterday’s tech leads! This change is reshaping what companies expect from their engineering leadership.

3. Interview process changes

Tech interviews are changing, and below are the biggest shifts from a few years ago that we’ve observed.

DSA interviews: elevated technical bar

On one hand, the fundamental structure of technical interviews hasn't radically changed. On the other, expectations have become significantly more demanding. Companies are simply setting a higher standard for what constitutes a passing performance.

In data structure and algorithm (DSA) interviews, engineers face noticeably harder problems at every stage of the process. One senior engineer interviewed at Google in 2021, and did so again last year. They told us:

"I used to think that LeetCode ‘hard’ problems were never asked at Google. Now [in 2024] they seem to have become the norm."

Beyond pure difficulty, we're seeing more emphasis on the completeness of implementation. Interviewers now routinely expect things like:

• proper error handling

• robust input validation

• clean code

• …all within the same time constraints as before.

There is little incentive to pass someone who doesn't get everything entirely correct. This is the grim reality of what happens when there are so many qualified candidates in the interview pool.

System design interviews: higher expectations

System design interviews have undergone an equally dramatic elevation. Senior-level candidates we talk to report being expected to demonstrate familiarity with modern distributed systems concepts that previously might have only been expected at staff levels.

Specialized knowledge has even crept into standard interviews. For example, geospatial indexing was once considered niche, but now has become commonplace in popular system design questions like "find nearby friends," Yelp-like applications, or ride-sharing platforms like Uber. We now advise candidates of all levels to have at least a basic familiarity with concepts like geohashing and spatial data structures (like quadtrees or R-trees) – as silly as that sounds. The same trends apply as to DSA: more candidates, more competition, a higher bar for hiring.

One staff engineer candidate we worked with really stood out. He had worked at Google in Seattle for almost 15 years, and was re-entering the market for the first time since. He was taken aback by the expectations in modern interviews compared to when he joined Google. As someone who had never before worked on stream processing systems, he found it frustrating that companies he interviewed at expected him to have intimate familiarity with concepts like exactly-once semantics, windowing techniques, and watermarking algorithms. He told us:

"I’ve built and maintained critical infrastructure for over a decade, but suddenly I'm expected to have specialized knowledge in areas completely unrelated to my expertise. It’s just so frustrating."

It's easy to empathize. At the same time, it's also easy to see how the luxury of choice with candidates leads to this. This elevation in technical expectations isn't arbitrary; with reduced hiring volumes, companies can afford to be more selective, and many are specifically looking for engineers who can contribute across a broader range of problems. Engineers with deep but narrow specialisms have fewer opportunities in this environment.

Downleveling

Downleveling seems to be a new trend. With heightened hiring bars and current market conditions, we're seeing candidates routinely receiving offers a level below their current position, particularly at the senior and staff levels.

In one case, Stefan worked with a candidate at Meta who successfully completed the interview process for a senior position later, but the offer was withdrawn and they were offered a mid-level role instead. This downleveling was due to a new policy requiring candidates to have at least six years of experience for senior positions. Personally, it’s heartbreaking – and arbitrary! – to see companies strongarm talent like this. The candidate ultimately accepted the offer, not being able to secure a better one.

This trend is particularly true for staff-level engineers, with many being offered senior positions even when they meet but don't easily exceed the staff-level bar. Companies have calculated that with less competition for talent, they can implement more aggressive leveling practices, and many candidates are accepting lower level offers after months of searching.

The long-term career implications of this are significant, as it often requires 2-3 years to get back to their former level. Despite this impact on career trajectory, we're seeing acceptance rates for down-leveled offers increase significantly as candidates prioritize stability in an uncertain market. Note from Gergely: we previously covered downleveling in The seniority rollercoaster.

Team match evolution

Perhaps the most significant structural change in the interview process has been the evolution of team matching. This is a process, now popular at Meta and Google, where candidates first pass an interview but don't receive offers until they match with a team.

This team matching approach has been adopted more broadly at larger tech companies, but with a slightly ugly twist: it's increasingly functioning as an additional filter, rather than for the candidate’s benefit.

We observe that team matching introduces a new set of "interviews" with hiring managers for candidates to navigate. It’s positioned as a mutual selection process, but the reality is that it's become another hurdle candidates must clear before securing an offer.

Meta notably overhauled its hiring process in 2024, eliminating most aspects of its longstanding "bootcamp" program, in which new hires joined the company first, and then found their team during bootcamp. In its place, they've implemented a team matching system that requires candidates to secure a team match before receiving a final offer.

The outcomes have been problematic for many candidates. One staff engineer we worked with who passed all technical rounds at Meta with strong, positive feedback, waited four months in team match limbo. To make things worse, by the time the team match completed, all their competing offers had expired!

When a match finally materialized, the offer was significantly below initial expectations, with little room for negotiation in the absence of alternatives. We see that team-matching backlogs seem to have been cleared as of late at Meta, but waiting many months remains common, especially in more competitive markets like New York City.

Indeed, some companies appear to be using team matching delays strategically as a negotiation tactic. Meanwhile, team-matching processes have morphed from giving candidates options, into additional screening layers where qualified candidates often find themselves eliminated or in limbo.

Team matching has evolved into a de facto second interview, despite companies' efforts to present it otherwise. From our conversations with hiring managers, we've found they commonly interview ten candidates to fill a single position. These managers strongly advise candidates to thoroughly prepare for this phase and customize their presentations specifically for the team they want to join.

Stefan advises candidates to plan for this phase and use it to their advantage. It’s true that the team matching process is slow – but this can create an opportunity to synchronize offers by scheduling interviews without team matching in place for later. Having several offers gives crucial leverage in negotiations.

4. Interview format differences at startups and Big Tech

With the rise of AI and growing skepticism about traditional coding interviews, we're seeing a widening gap between how Big Tech and newer companies do interviews.

Traditional FAANG employers remain largely committed to their existing formats, with only minor adjustments. As one FAANG head of recruiting told us:

"The inertia of these processes is enormous. These companies have built entire recruiting machines around their current processes, with years of calibration data. They're reluctant to make dramatic changes without compelling evidence that alternatives would work better at scale."

Organizationally, changes to the interview process are often gatekept by engineering executives who would prefer to wait for a fire, than potentially create a problem at the first smell of smoke.

Several mid-sized companies have moved toward more realistic, open-ended coding challenges that better reflect actual work. Examples of places adopting more realistic interviews include Stripe, Coinbase, and OpenAI. Rather than solving LeetCode questions, candidates tackle problems like:

• Designing a query engine

• Implementing a key-value store

• Designing an in-memory database to handle transactions

Early-stage startups have pushed even further, often replacing traditional coding exercises with take-home projects that explicitly allow the use of AI tools. Yangshun Tay, founder of GreatFrontEnd, has been a prominent voice on Linkedin advocating for this shift in hiring practices. He detailed how his team successfully implemented this approach to better evaluate candidates' real-world problem-solving abilities:

“Coming from Big Tech, I'm aware of the flaws of the typical interview process. Hence I use a somewhat different process when it comes to hiring Front End Engineers for GreatFrontEnd:

1. Zero LeetCode

2. Take-home assignment

3. The take-home assignment is a todo list (what?!)

4. Product sense is evaluated

5. Candidates who pass the take-home assignment know the upcoming interview questions beforehand and have ample time to prepare

6. Candidates get a perk for interviewing with us (...)

It's important to note that such an interview process is more time consuming than the standard LeetCode one and does not scale well with the number of applicants.”

This shift serves a dual purpose: it better reflects real work conditions, while combating the growing problem of assessment fraud. One seed-stage AI founder in the Bay Area we spoke with estimated that at least 20% of candidates were obviously cheating in their traditional coding tests. The issue isn't limited to startups; one of Evan’s good friends, an Amazon interviewer, confided that half of his last ten candidates were obviously using AI tools on secondary screens during supposedly monitored assessments. By explicitly incorporating these tools into the evaluation process, companies are adapting to workplace realities and assessments’ integrity challenges.

Innovation in technical evaluation is bubbling up from smaller, more agile organizations, with Big Tech watching on from behind. This is an interesting inversion of the historical pattern wherein for the past decade, interview practices pioneered by Google and other tech giants trickled down to smaller companies eager to emulate their success. Now, it’s the opposite! One question is when or if FAANG employers will adapt to this new reality.

The truth is that Big Tech is unlikely to make changes to the hiring process without resounding, negative post-interview signals, which could be things like a significant quantity of unregretted attrition attributable to poor interview signals.

We think it’s more likely that Big Tech makes minor adjustments, like returning to on-site interviews in the short term. They recognize their current interview processes are essentially a game, but they do effectively identify candidates willing to invest in intensive preparation. Unfortunately for candidates, their willingness to grind through arbitrary algorithmic challenges correlates just enough with on-the-job characteristics of high-performing engineers to justify maintaining the status quo.

We wonder if sticking to existing interview approaches is increasingly unsustainable in the age of large language models (LLMs). As AI tools become more capable of solving the exact algorithmic puzzles used in interviews, the signal value of those assessments will inevitably diminish. No engineer in the future will need to manually code algorithms like parenthesis balancing or binary tree traversals; instead, they'll prompt an AI to generate that code. The companies pioneering more realistic, project-based assessments are adapting to the reality of how engineering work will actually be done moving forward.

What's clear is that candidates currently face a bifurcated landscape: prepare for traditional algorithm interviews for Big Tech roles, while simultaneously developing the skills to excel in more open-ended, practical evaluations for opportunities elsewhere.

5. Preparation strategies by experience level

We’ve found that an optimal preparation strategy varies significantly by experience level, and the relative importance of different interview components change with career progress. Here are patterns we’ve observed:

For junior engineers with 0-2 years of experience, we’ve found this preparation the most effective:

• 80% of preparation time: focus on algorithms and coding problems

• 20%: preparation for the behavioral interviews

The technical bar for junior roles has risen dramatically, making mastery of fundamental algorithms and data structures essential. Successful junior candidates typically solve 150-200 coding problems across all difficulty levels before interviewing. You must be a stronger coder before anything else.

Mid-level engineers with 2-4 years of experience benefit from a more balanced approach:

• 50% coding

• 25% system design

• 25% preparation for behavioral interviews

At this level, companies expect strong implementation skills and emergent architectural thinking. The most successful mid-level candidates we work with develop a systematic approach to system design, focusing on building blocks they can combine and adapt, rather than memorizing specific solutions.

For senior engineers with 5-8 years’ experience, we’ve seen this setup work well:

• 50% preparation on system design

• 20% on coding

• 30% on behavioral interviews

The primary differentiator at this level is the ability to design robust, scalable systems while clearly articulating tradeoffs. Senior engineers are expected to handle ambiguity well, asking clarifying questions and making reasonable assumptions when information is incomplete.

The most common mistake we see from senior candidates is neglecting behavioral preparation. This is a critical error; at the senior level, companies are evaluating not just technical capability, but also leadership potential, conflict resolution skills, and cultural fit.

We've seen technically brilliant candidates fail interviews or get down-leveled unnecessarily because they couldn't effectively communicate their impact, describe how they influenced cross-functional teams, or demonstrate self-awareness about previous challenges. Behavioral preparation isn't a checkbox; it significantly impacts hiring decisions, especially at senior levels and above.

Staff+ engineers face a different challenge:

• Coding: a baseline at this point; stumble here and rejection can be swift.

• 90% of differentiation comes from system design and behavioral/leadership assessments.

For these roles, companies look beyond implementation details to evaluate architectural vision, cross-functional leadership, and executive communication skills. Successful staff+ candidates demonstrate strategic thinking, connecting technical decisions to business outcomes in their system design discussions.

Top AI labs like OpenAI have their own distinct hiring patterns. Rather than prioritizing traditional leadership skills, they heavily filter by pedigree or headline achievements and strongly favor candidates from elite, high profile companies, AI-focused startups, prestigious universities, and those with flashy achievements which they can communicate easily. Without this, applicants face an uphill battle, regardless of their technical excellence.

Effective Practice

Let's acknowledge the reality that the tech interview process has become a specialized game that continues to deviate from day-to-day engineering work. This isn't ideal, but it's the reality. Companies have settled on standardized evaluation approaches that don't perfectly mirror actual job responsibilities, and this disconnect frustrates many engineers.

The good news is that the rules of the game are publicly known. It's essentially a “secret handshake” you need to learn to gain entry into these companies. The process might seem arbitrary, but with proper preparation, it's entirely learnable. Anyone with sufficient dedication can master these patterns and significantly improve their performance.

We recognize our bias here; as an interview preparation platform, we obviously believe in the value of structured practice. The data speaks for itself: candidates who engage in deliberate practice consistently outperform those who don't, regardless of natural ability or experience level. The patterns are clear across thousands of interview outcomes.

If investing in formal mock interviews doesn't fit your preferences or budget, that's completely understandable and there are numerous alternatives: find a friend who works at your target company, connect with peers on Reddit or Discord communities, or form study groups with other job seekers. The specific method matters less than the fundamental principle that interviewing is a skill that improves with practice, feedback, and iteration.

What doesn't work is assuming your daily engineering work has prepared you for the interview environment. The performance aspect of interviewing – thinking aloud, handling pressure, communicating clearly while solving problems – requires deliberate practice in conditions that mirror actual interviews. Without this, even brilliant engineers can struggle to demonstrate their capabilities within the artificial constraints of an interview process.

6. Silver linings

The tech hiring landscape of 2025 is a far cry from the job seekers’ gold rush of 2020-2021. The pendulum has swung hard from "please take our money" to "prove you're worth it," and engineers are feeling the squeeze. But don't throw in the towel!

Major tech companies including Amazon, Apple, Microsoft, Google, and Meta, collectively maintain nearly 40,000 open roles. Even orgs inside these companies that aren't growing headcount-wise are still hiring to backfill, even as they make layoffs.

The AI sector continues to experience exceptional growth, with companies like OpenAI, Anthropic, and numerous AI infrastructure startups, hiring aggressively. Ignore bleak predictions about the imminent replacement of engineers; the reality of hiring shows how much businesses need engineers to achieve their goals. Companies in the AI sector often offer comp packages reminiscent of the 2021 peak, particularly for engineers with relevant expertise, or those demonstrating strong learning potential in AI-adjacent domains.

The engineers crushing it right now understand that modern tech interviewing has basically become its own bizarre sport, complete with arbitrary rules and peculiar traditions. They're treating interviews like performances, not just technical evaluations, and they're putting in the practice hours to match.

As AI continues reshaping how engineering work gets done, interview processes will have to evolve too because they can't keep testing for skills which AI makes obsolete. But for now, we're stuck in this awkward transitional phase where you need to master both the old-school algorithm games, and newer, more practical assessments.

At the risk of stating the obvious, at Hello Interview we see a clear pattern: there's a strong correlation between preparation investment and interview success. Candidates who dedicate time to structured practice are still more likely to secure multiple offers, even in this more selective environment.

The game might be harder, but at least the rulebook is public. With enough deliberate practice and the right preparation strategy for your level, you can still come out on top, even in this tougher market.

Takeaways

Gergely again.

Thanks very much to Evan and Stefan for summarizing what they see in the tech market. They gathered most of these insights by interacting with devs using Hello Interview, the mock interview and interview prep service of which they’re cofounders.

Some interesting things I’m reflecting on:

A tighter job market was predicted due to the end of zero interest rates. A year ago, we analyzed what the end of 0% interest rates meant for software engineers. One top conclusion was to expect a tougher job market: more competition, less “shopping around”, and engineering managers especially having a hard time. This is reflected in what Evan and Stefan report from the recruitment front line. In some ways, since this was a predictable effect, it’s easier to understand as a cause of some of today’s hiring challenges.

AI is hot, and hot AI companies care a lot about pedigree. One thing that feels a little surprising is learning from Evan and Stefan how much companies like OpenAI, Anthropic, and other top companies seem to filter by pedigree. Perhaps this should not be a surprise given these companies are inundated with applications from everywhere: if they can afford to hire from the most glittering tech companies, why wouldn’t they?

It’s a reminder that even if you are in the AI engineering field, getting into a top workplace from a lesser-known company is close to impossible without recognised industry expertise. If your target is to one day work at such companies, you might need to prepare for a multi-step career journey, starting out at lesser-known places doing AI-related work, forging a path towards better known companies, and then, one day, perhaps getting into the “top” places.

Getting a job takes more time investment. There’s constant complaints from devs about how time-consuming it is to prepare for tech interviews. If it’s a company doing Leetcode-style algorithmic interviews, the issue is the time it takes to practise beforehand. Meanwhile, if it’s a company with a complex takehome, the complaint is the time it takes to do the exercise itself.

As a competent engineer, you probably assume that an employer should accept that you know your craft, skip all the time-consuming evaluation that just seems pointless, and lead with an offer, right?

But from the vantage point of a hiring manager, a new hire is always a risk. Few things are worse than a hire that doesn’t work out because of a skill gap, motivation, or any number of reasons. Interviews are meant to verify skillsets and motivation.

We’ve gone from a candidate’s market in 2020-2022, to an employer’s market today. In this environment, you will most likely need to invest a lot more time in preparing for job interviews and doing them. The upside is that this preparation doesn’t go to waste, as interview formats don’t all change so rapidly.

Expectations are going up, and will keep on rising. At Uber, my manager at the time told me that performance expectations at Uber only went one way for the same level: up. This was because the business was growing rapidly, and the expectation of any new hire was to raise the bar. This meant that over time, expectations of “normal” for any career level kept inflating. After a while, this felt natural, but it was a strange thing to adjust to!

I feel we’re seeing something similar play out across the broader industry, today. Due to lots of qualified, capable engineers applying for jobs, expectations are going up at all career levels, and this is why downleveling is more common.

If you get a downleveled offer: first of all, congrats for getting an offer! In this job market, it’s an achievement in itself. It’s helpful to take current job market conditions into account before being too disappointed by this outcome. And if you have yet to receive an offer, know that the market is tougher right now than it’s been in a decade, so job searches take longer than before.

If you’re currently in the job market, or are doing interviews and have noticed any changes in recruitment processes, please share in a comment:

This deepdive is a guest post – if you’re interested in pitching an article for The Pragmatic Engineer, see details here.