Collinear powers leading AI labs and Fortune 500 enterprises with high-signal post-training datasets and RL environments, helping teams improve model quality faster and at lower cost.
We simulate realistic, multi-turn user and adversarial scenarios to uncover how models fail in production, then turn those failures into curated data and reinforcement learning environments for CPT, SFT, and RL. The result is faster convergence, stronger robustness, and measurable gains where it matters most.
Own real research direction: Set the agenda for mid-training and post-training data curation, agent simulations, and evaluation in a production-facing system.
Translate research into impact: See your work directly influence customer outcomes, not just academic benchmarks.
Work at the core of LLM quality: Agent gym, scalable infrastructure, multi-agent systems, context management, , eval design, reward modeling, CPT/SFT/RL data.
Small, high-trust team: You get to lead a strong research team with PhDs from top universities, Tight feedback loops, fast iteration, and direct access to founders.
Strong footing for growth: VC-backed in the heart of silicon valley, active enterprise customers, and room to shape the company’s technical identity.
Customer impact: You will play a critical role of taking vague customer problems and translating into research roadmap for the team
As Head of Research, you will own Collinear’s research vision and execution across evaluation, simulation environments, and post-training. This is a hands-on leadership role at the intersection of research, product, and customer impact.
You will lead a small, high-caliber team of research scientists and engineers, while working closely with founders and customers to ensure our research translates into measurable improvements in real-world AI systems.
This role is ideal for someone who wants to define a research agenda, build a team around it, and see their work shipped into production.
What You’ll DoDefine and lead Collinear’s research roadmap across:
LLM evaluation and benchmark design
Agent-based and multi-turn simulation environments
Post-training methods (CPT, SFT, RLHF/RLAIF, reward modeling)
High-signal data and semi-synthetic data generation
Lead and grow a team of research scientists and research/ML engineers.
Translate research ideas into scalable, production-ready systems.
Partner closely with founders on technical strategy, roadmap, and prioritization.
Collaborate with enterprise customers to understand real failure modes and guide research direction.Represent Collinear externally through papers, talks, open-source contributions, or community engagement.
We’re looking for someone with a strong research foundation and the ability to lead and execute in an applied, fast-moving environment.
You likely have:
PhD in CS or related area
7–10+ years of experience in ML research, applied research, or research engineering.
Experience leading research projects or teams (formal management experience is a plus but not required).
Deep familiarity with LLMs and post-training techniques, such as:
Evaluation and benchmark design
RLHF / RLAIF / reward modeling
Synthetic or semi-synthetic data generation
Red-teaming, robustness, or failure-mode analysis
Experience publishing at or contributing to top-tier venues (e.g., NeurIPS, ICML, ICLR, ACL, EMNLP) and/or impactful open-source projects.
Strong communication skills and comfort working across research, engineering, and product.
Experience in an industry research lab or fast-growing startup.
Prior work on agent systems, interactive environments, or multi-turn evaluation.
Experience working directly with customers or external stakeholders.
Comfort operating with ambiguity and high ownership at an early-stage company.
Collinear is an equal opportunity employer and values diversity. We do not discriminate on the basis of race, color, religion, sex, gender identity, sexual orientation, national origin, age, disability, veteran status, or any other characteristic protected by applicable law.
The base salary range for this role in California is $250,000 to $400,000 per year, depending on experience, skills, and qualifications. This role will also be eligible for equity, benefits, and bonuses.
Collinear provides reasonable accommodations for candidates with disabilities throughout the application and hiring process. If you need an accommodation, please contact us.
Pursuant to applicable local ordinances, we will consider qualified applicants with arrest and conviction records.
Skills Required
- PhD in Computer Science or related area
- 7-10+ years of experience in ML research or applied research
- Experience leading research projects or teams
- Deep familiarity with LLMs and post-training techniques
- Experience publishing at top-tier conferences
What We Do
Collinear AI builds simulation labs where AI agents learn to work in the real world by simulating users, tools, and workflows to improve AI models before deployment, focusing on AI safety, reliability, and customization for enterprise GenAI.







