AI & ML Engineer Recruiting Firm: How to Hire Machine Learning Engineers in 2026

By Recruiting from Scratch · Updated monthly · Last updated September 2026

Data: live machine learning and AI engineering postings at 33 AI labs, AI startups and big tech companies tracked by Recruiting from Scratch (deduplicated by company and title, internships excluded) + U.S. Department of Labor H-1B wage filings, FY2025–FY2026. All figures are base salary; equity and bonus are paid on top.

Hiring machine learning engineers in 2026 means competing with AI labs that post the highest base salaries in software, big tech companies with the most open roles, and well-funded AI startups offering equity upside. The engineers who can train, fine-tune, evaluate and serve models in production are the most contested hires in the market — and most of them are already employed.

Recruiting from Scratch is a technical recruiting firm that places ML engineers, AI engineers, research engineers and the infrastructure engineers behind them. We track every ML and AI engineering posting at the companies competing for this talent, and the wages they file with the Department of Labor, so hiring managers know what the market pays before they start a search.

Who is hiring machine learning and AI engineers right now

Across the companies we track there are 1,418 distinct ML and AI engineering roles open: 998 at big tech, 257 at AI labs and 163 at 21 AI startups. The most active hirers:

Among AI labs, OpenAI has 75 open ML and AI engineering roles, Anthropic 74 roles and Perplexity 40 roles. Among AI startups, Databricks has 21 roles, Abridge 19 roles and Mercor 18 roles. At big tech, Amazon lists 348 roles and Google 200 roles.

What this means: big tech still has the most ML roles by volume, but AI labs post the highest pay ranges as a group (Netflix is the notable big-tech exception, with wide posted ranges), and AI startups are where the competition for senior ML engineers is fiercest — they need the same people with far smaller teams. If you are hiring your first or fifth ML engineer, you are competing against every company in this table.

What machine learning engineers are paid in 2026 (base salary)

AI labs set the ceiling on base. Big tech pairs a lower base with RSUs and refreshers. AI startups sit between the two on base and compete with equity. A startup band set against big-tech numbers will lose ML candidates to the labs; a band set against the labs is often out of reach — which is why the equity story and the problem matter as much as cash.

How to hire machine learning engineers: what actually works

• Define the ML role precisely. "ML engineer" covers model training, applied ML in product, ML infrastructure, evaluation and inference optimization. Candidates self-select on the actual work — write the job around the model and the data, not a generic list of frameworks.
• Benchmark pay before you post. Use the ranges above. Most failed ML searches fail at the offer stage, not the sourcing stage.
• Source from where ML engineers already are. The strongest candidates are employed at the companies in the table above and rarely apply to postings. Direct outreach with a specific technical pitch outperforms job boards.
• Run a short, technical loop. One practical ML exercise plus one systems conversation beats five generic interviews. Long, multi-week loops are where companies lose ML candidates to faster offers.
• Sell the problem and the data. ML engineers choose employers on the problem, the data they will have access to, and the compute they can use — then on pay.

How long does it take to hire a machine learning engineer? The industry benchmark is about 49 days from search start to hire. Recruiting from Scratch averages 29 days to hire across 300+ engineering placements.

What candidates should know about ML offers in 2026

• Compare the whole package. AI labs lead on base; big tech adds RSUs and refreshers; startups trade base for equity. Model four years, not one.
• Ask what you will actually work on. "ML engineer" can mean research, product ML, infrastructure or evaluation — the day-to-day varies more than the title.
• Ask about compute and data access. They decide what you can build, and they differ widely between companies.
• Posted ranges are a floor. Pay-transparency laws require a range; the filed wages above are a better guide to what companies actually pay.

Before you start a search: an ML hiring readiness checklist

• A pay band checked against AI-lab and startup ranges, not only big tech.
• A clear answer to "what model, what data, what compute?"
• A hiring manager who can run the technical interview.
• An interview loop that fits inside two weeks.
• A decision on location: many ML candidates hold remote or multi-city offers.

The ML and AI roles we recruit

• Machine learning engineers (applied ML, recommendation, ranking)
• AI engineers building LLM-powered products and agents
• Research engineers and research scientists
• ML infrastructure, training and inference engineers
• Forward-deployed and applied AI engineers
• Data engineers for ML platforms

Recruiting from Scratch has recruited engineers for AI companies including Mercor and Decagon.

Top AI & ML engineer recruiting firms (2026)

For individual-contributor ML engineers, choose a firm that recruits engineers every day — not only executives — and can benchmark AI-lab, startup and big-tech offers side by side.

Why hiring managers work with Recruiting from Scratch

• 300+ engineering placements and 150+ companies served
• 29-day average time to hire, versus a 49-day industry benchmark
• A network of 2M+ candidates and pay data from 1M+ job postings and federal filings

Hiring machine learning or AI engineers? Book a call with our team or tell us about the role.

An ML engineer looking at AI labs and startups? See open roles and share your resume.

More: AI Engineer Recruiting Firms in San Francisco · Hedge Funds vs AI Labs vs Big Tech: Engineer Pay · Quant & Hedge Fund Recruiting in New York

Methodology

• Job postings: active ML and AI engineering postings (titles such as machine learning engineer, AI engineer, research engineer, research scientist, member of technical staff, inference and training engineer) at the companies named on this page, from their public job boards. Each role is counted once per company and title. Internships, campus programs, managers and analysts are excluded.
• Pay: U.S. Department of Labor H-1B filings with the same title families, fiscal years 2025–2026, certified, annual wages; median of the filed low and high ends. Segments with fewer than 5 filings are not shown.
• What's not included: bonus, equity and benefits. Companies appear as market data only; apart from the clients named above, their inclusion does not mean they are Recruiting from Scratch clients.
• Refresh: monthly.

FAQ

How long does it take to hire a machine learning engineer?
About 49 days on average across the industry. Recruiting from Scratch averages 29 days to hire, because most of the time in an ML search is lost to unclear role definitions, off-market pay bands and long interview loops.

How much do machine learning engineers make at AI labs?
H-1B filings for AI-lab ML and AI engineering roles (FY2025–FY2026) show a median filed base of $300K–$335K, with the 90th percentile at $460K. Equity is paid on top.

Which companies are hiring the most machine learning engineers in 2026?
In our current tracking, big tech has the most open ML roles, AI labs pay the highest base, and AI startups are hiring aggressively for smaller, senior teams. See the table above for the most active companies.

What is the best recruiting firm for hiring ML engineers?
Look for a firm that recruits ML engineers every day, knows what AI labs, startups and big tech are each offering your candidates, and can show its time to hire. Recruiting from Scratch averages 29 days to hire across 300+ engineering placements.