Mechanize
Builds reinforcement learning environments and evals that score frontier coding agents on software engineering tasks
Highlights
- New on EvolcAdded 2026-11-13
About Mechanize
Mechanize, Inc. builds environments and evaluations for frontier coding agents. In these environments, models carry out software engineering work such as building a feature, deploying an application, or debugging an issue in an unfamiliar codebase. A grader scores the model's performance, and the scores serve as signals during reinforcement learning and evaluations.
The company publishes GBA Eval, a benchmark that measures how well coding agents can write a Game Boy Advance emulator from scratch in 24 hours. Its founders are Matthew Barnett, Tamay Besiroglu and Ege Erdil. Backers named on the site include Nat Friedman and Daniel Gross, Patrick Collison, Adam D'Angelo, Marco Mascorro, Dwarkesh Patel, Sholto Douglas, Devendra Chaplot, Alex Atallah and Marcus Abramovitch.
Mechanize is hiring software engineers to design and build these environments, with open roles at intern, junior, mid-level and senior levels in San Francisco.
Product: Reinforcement learning environments and graded evaluations for frontier coding agents, plus the GBA Eval benchmark.
Jobs at Mechanize
More ai & ml companies
Mercor
Provides expert-generated training data, AI evaluations and custom agents for AI labs and enterprisesYoung companySalary transparentMeshy
AI 3D creation tools for game developers, 3D printing users, designers, and creative teamsNew on EvolcMetafore
Enterprise AI orchestration software that connects company data and automates workflows for large organizationsYoung companyRemote-friendlyMia Labs
AI voice and text agent that answers calls, books service appointments and follows up with customers for car dealershipsNew on EvolcMirage
AI video creation app, agent-ready creative tools and video APIs for creators, businesses and developersYoung companyNew on EvolcMirendil
Frontier AI models and autonomous research systems for AI labs and science-focused organizationsNew on Evolc
Website technologies
Tools and platforms detected on this company’s website, grouped by purpose. Dates record when Evolc observed them.
Analytics
- Google Analytics 4High confidence
- First detected
- Last detected
- Last checked
Detection evidence
- Website resourcedeclared script/iframe or software asset
- PostHogMedium confidence
- First detected
- Last detected
- Last checked
Detection evidence
- Integration code on pageinitialization/configuration expression
Content delivery
- CloudflareHigh confidence
- First detected
- Last detected
- Last checked
Detection evidence
- Response headerserver
- Response headerserver
Hosting
- AWSHigh confidence
- First detected
- Last detected
- Last checked
Detection evidence
- Response headerx-amz-cf-id
- Response headerx-amz-cf-id
Marketing & advertising
- Google AdsMedium confidence
- First detected
- Last detected
- Last checked
Detection evidence
- Integration code on pageinitialization/configuration expression
- LinkedIn InsightMedium confidence
- First detected
- Last detected
- Last checked
Detection evidence
- Integration code on pageinline script loader
- X PixelMedium confidence
- First detected
- Last detected
- Last checked
Detection evidence
- Integration code on pageinline script loader
Website detection does not establish a paid subscription or company-wide use. Tools loaded after consent or sign-in may be missed. “No longer observed” does not prove the company stopped using a tool.