Hiring great machine learning talent isn’t about posting on Hacker News and waiting. In 2025, the best engineers are more intentional, more ambitious, and more selective than ever before.
With San Francisco continuing to dominate as the global hub for AI startups, the competition isn’t just about compensation anymore. It’s about meaning, impact, vision, autonomy, talent density, and your story.
At Codexo, we speak daily with the ML Engineers and Applied Researchers who power early-stage innovation. Here’s what they’re actually looking for and how the best startups are adapting to attract them.
1. Day 1 impact: Engineers want to solve real problems, fast
It’s no secret that the best engineers want their work to matter. But it’s the type of impact they care about that’s shifted.
They don’t want to ship pipelines no one uses, or clean data for someone else’s OKRs. They want to work on applied, challenging problems where their contributions are visible early – often in the first month.
Think:
-
Building a production-ready recommendation engine from scratch
-
Designing experiments that feed directly into product decisions
-
Shaping a new research direction with real user impact
They’re not interested in “tech for tech’s sake.” They’re looking for problems grounded in use cases, real users, and ambitious outcomes.
But why does this matter to founders? You’re competing not just with Big Tech pay, but with engineers’ desire for creative autonomy and real-world impact from the start.
According to the 2025 AI & Machine Learning Salary Outlook, ML Engineers in San Francisco command average total compensation around $193,000+, significantly higher than typical Data Engineer roles. That premium reflects both the technical complexity and the real business impact they’re expected to deliver.
And Business Insider recently reported that many smaller tech firms, especially startups, are paying premium bonuses up to $200,000 for professionals with ML/AI experience, underscoring how fiercely competitive this talent market has become.
Together, these trends suggest that cash isn’t the decider, impact is. If you sell the level of problem exposure and strategic ownership an ML Engineer will have from day one, you’re already playing the right game.
2. Working directly with technical founders is a differentiator
The calibre and accessibility of your founding team matters, especially in ML hiring.
We consistently hear from engineers that working directly with founders is a major draw, particularly when the founders bring:
-
A strong academic background (PhD-level research, open-source contributions, published papers)
-
Credibility in the space (previous exits, prior roles at DeepMind, OpenAI, Meta AI etc.)
-
Technical alignment – someone who understands what it actually takes to solve these problems
This isn’t about ego. It’s about trust.
Engineers want to work closely with people who are in the trenches – shaping the product roadmap, reviewing architecture, and solving problems side by side.
The “you’ll report to the VP of Engineering” line doesn’t hit the same if the founder is completely disconnected from the tech.
Founders don’t have to code every day, but they do need to be intellectually and strategically engaged with the work. That’s how trust and talent magnetism is built.
3. Talent density: joining a team they can learn from
If there’s one through-line across every strong ML hire we’ve supported, it’s this: great engineers want to be surrounded by people they can learn from.
Not in theory, in practice. In Stack Overflow’s 2024 Developer Survey, 42% of ML respondents cited “opportunities to learn” as their top reason for considering a new role, more than compensation or job title.
ML candidates are looking at the backgrounds of your current team, and asking who they’ll pair with and what that person has shipped. It’s not enough to say “we’re smart.” You need to show it.
For startups, it’s crucial to highlight your team’s credentials early in the hiring process. If your first engineer used to work on robotics at Google Research or helped scale infra at Anthropic, say it.
And if you don’t yet have that kind of talent density? Be honest about that too and tell a compelling story about why they’ll be the one to set the bar.
4. A culture of autonomy, not bureaucracy
One of the most common frustrations we hear from engineers exiting big tech is the glacial pace of progress.
Endless approval layers, red tape, legacy systems, and decisions made three levels above them.
That’s your opportunity.
Startups have a window to offer something Big Tech can’t: real autonomy.
Engineers want to move fast, test assumptions, ship experiments, and iterate without friction. Especially in ML environments where iteration speed is a competitive advantage.
That doesn’t mean chaos, it means trust.
The ability to own a problem space. The freedom to shape architecture. The mandate to propose new directions without navigating a six-step review process.
5. A compelling mission and urgency around it
A generic “we’re building tools for better decision-making using AI” won’t cut it.
In 2025, every startup is an AI startup. Every company is “solving a real problem.” Your mission needs to cut through the noise.
Engineers want to know:
-
What’s the real problem you’re solving?
-
Who’s it for?
-
Why now?
-
Why does it matter?
They’re not just scanning for compensation and equity – they want a reason to care.
And more than that, they want to feel momentum. The sense that things are moving. That your company isn’t just promising a big vision, but actually taking steps to realise it.
Because smart engineers are also strategic. They want to join rocket ships, not deckchairs.
Final thoughts
You don’t need to tick every box, but you do need to know which ones you can.
You might not have Series A funding yet. You might not have a full ML team. But what you can do is show self-awareness, and build a compelling offer around the strengths you do have.
Hiring great ML engineers is as much about storytelling as it is about sourcing. The best candidates don’t just want a job. They want a journey.
Tell a better one, and they’ll follow.
Need help hiring ML engineers in San Francisco?
Codexo works with the most ambitious startups in the AI space. We partner with founders to attract, engage, and close top-tier Machine Learning talent who move the needle.
Want to see how we’d approach your hiring strategy? Get in touch with us today.