Pay
$166K–$237K is typical
The middle half of what similar US postings state. This one doesn’t list pay.
From 1,537 postings in the last year. Stated ranges, not salaries paid.
San Francisco · San Francisco, California, United States
Open 5 months, first seen by us 172 days agoLocation type: On-siteFrom the employer’s posting
Employment type: Full-timeFrom the employer’s posting
Department: EngineeringFrom the employer’s posting
We're looking for founding Machine Learning Engineers (MLEs) to own and improve our core action models end-to-end - the intelligence that powers Composite's proactive automation platform.
You'll work at the intersection of LLM inference, browser understanding, and low-latency systems, shipping models that need to feel instant while reasoning over complex page state and user context.
Improve accuracy and latency of core models predicting users' next actions
Design and optimize LLM inference pipelines, including caching and streaming
Build evaluation frameworks and data pipelines to measure model quality
Develop synthetic data generation pipelines for browser interaction training
Ship features end-to-end that go directly to users
We're looking for founding Machine Learning Engineers (MLEs) to own and improve our core action models end-to-end - the intelligence that powers Composite's proactive automation platform.
You'll work at the intersection of LLM inference, browser understanding, and low-latency systems, shipping models that need to feel instant while reasoning over complex page state and user context.
Improve the accuracy and latency of our core models across diverse web applications to predict users' intended next actions
Design and optimize LLM inference pipelines, including token caching strategies, streaming architectures, and network-level optimizations between client and server
Build evaluation frameworks and data pipelines to measure and improve model quality at scale
Develop synthetic data generation pipelines for browser interaction training data
Ship features end-to-end that go directly to users — this is not a research-only role
College-educated professionals spend 85% of their day as digital factory workers in Chrome, clicking through repetitive browser tasks.
In the employer’s words · Read in context · Company website ↗
Open 5 monthsWe first saw this posting 172 days ago and it is still up. Roles open this long are often slow to fill or kept open to collect applications.
These dates show when we found the posting. Check the employer’s website to confirm it is still accepting applications.
$166K–$237K is typical
The middle half of what similar US postings state. This one doesn’t list pay.
From 1,537 postings in the last year. Stated ranges, not salaries paid.
Day 172
Up longer than most. Similar jobs usually come down between day 7 and day 76.
From 6,398 similar postings that came down. Coming down is not always a hire.
1 open job
Add your resume and we’ll mark which of these tools you already use, and point out anything in the posting that could rule you out.
- Experience with vector databases (e.g., Turbopuffer, Pinecone) and hybrid search - Full-stack development experience (TypeScript, Node.js, React) Our Values
Complete your application on jobs.ashbyhq.com. The employer’s form will show what is required.
Already applied? Track this application
We're looking for founding Machine Learning Engineers (MLEs) to own and improve our core action models end-to-end - the intelligence that powers Composite's proactive automation platform.
You'll work at the intersection of LLM inference, browser understanding, and low-latency systems, shipping models that need to feel instant while reasoning over complex page state and user context.
Unlike hosted browser solutions that introduce latency and auth barriers, or consumer-focused "AI browsers," we run AI directly through professionals' existing browsers via a Chrome extension, creating instant response times with zero migration or IT friction. This architecture creates unique ML challenges.
This is a high-ownership role on our small, exceptional team where your work ships directly to users and has the potential to tangibly improve the work lives of hundreds of millions of people.
About CompositeCollege-educated professionals spend 85% of their day as digital factory workers in Chrome, clicking through repetitive browser tasks.
Composite is building the proactive layer for productivity so professionals around the world can focus on meaningful, high-leverage work. We're training action prediction models that run in real time, anticipating what you'll do next based on page context and prior interactions.
We've raised $5.6M in seed funding led by Nat Friedman and Daniel Gross, with participation from Menlo Ventures, Anthropic's Anthology Fund, SVAngel, and other incredible investors.
What You'll Work OnImprove the accuracy and latency of our core models across diverse web applications to predict users' intended next actions and execute them faster than manual input
Design and optimize LLM inference pipelines, including token caching strategies, streaming architectures, and network-level optimizations between client and server
Build evaluation frameworks and data pipelines to measure and improve model quality at scale
Experiment with retrieval-augmented approaches using vector databases for contextual memory
Develop synthetic data generation pipelines for browser interaction training data
Work with DOM states, accessibility trees, and user interaction data to improve browser understanding
Ship features end-to-end that go directly to users — this is not a research-only role
Strong ML fundamentals with hands-on experience training and deploying models in production
Obsessive about latency — experience optimizing inference pipelines to feel instant to end users
Deep care about data quality, with the instinct to build tooling that ensures it
Experience with LLMs, transformer architectures, or sequence prediction problems
Comfortable working across the stack — our system spans a Chrome extension, Electron app, Cloudflare Workers edge proxy, and inference providers
Character: You're someone we'd want to work closely with for the next ten years. You approach challenges with curiosity rather than ego. You're a team player, a great communicator, and aren't afraid to be wrong.
Work Ethic: You're energized by hard problems and comfortable working intensely toward ambitious goals.
Raw Intelligence: You can quickly understand complex systems and solve novel, ambiguous problems with self-guidance.
Experience with browser automation, Chrome extensions, or web scraping at scale
Familiarity with accessibility tree / DOM parsing for page understanding
Background in RL or online learning from user interaction data
Experience with vector databases (e.g., Turbopuffer, Pinecone) and hybrid search
Full-stack development experience (TypeScript, Node.js, React)
Disagree and commit: Respectfully challenge decisions you disagree with, even when it's uncomfortable. Don't censor yourself or your ideas. Once a decision is determined, everyone commits wholly.
Clear and consistent standards: Decisions are made based on a shared framework that applies for everyone. We don't leave room for "rules for you, not for me" or any perceived hypocrisy.
Over-communicate: Nothing slows down a company more than confusion, mis-, or under-communication. Leave no room for ambiguity. Ask dumb questions. Write things down clearly.
Health is #1: Stay hydrated. Eat a balanced diet. Sleep 8 hours a night. Exercise frequently. Maintain good social and mental health. Not doing so affects your mood and long-term productivity.
Do the right thing.
New ML Engineer postings at the employers we follow are down 4% over the last 6 weeks compared with the 6 before.
Each bar is one week of new postings. The darkest bars are the last 6 weeks.
Counted from the hiring pages we’ve followed for at least five weeks, so employers we add don’t look like growth. Updated weekly.
We file this job under ML Engineer, by the work the posting describes.
ML engineers put machine learning models into products. They write the code that trains, ships, serves and monitors models so a recommendation or prediction works for real users.
Postings ask for strong Python and software engineering, a framework such as PyTorch, cloud services, and more and more often experience with large language models.
A data scientist often works out which model to build; an ML engineer makes it run reliably in production.
The government has no code for ML engineers, so these figures are for the closest group, computer and information research scientists.
Source: US Bureau of Labor Statistics, Occupational Employment and Wage Statistics (May 2025) and Employment Projections (2025 to 2035). Occupation code 15-1221.
Selected passages from the saved posting. Check the full description for conditions and exceptions.
No pay amount identified in the saved description.
San Francisco, California, United States
- Background in RL or online learning from user interaction data - Experience with vector databases (e.g., Turbopuffer, Pinecone) and hybrid search - Full-stack development experience (TypeScript, Node.js, React)
No clear work-authorization passage found. Eligibility is unconfirmed.