Forward Deployed Research Scientist (AI)

TL;DR

Forward Deployed Research Scientist (AI): Engage directly with frontier AI labs on data strategies and model fine-tuning, run ablation studies on open-weight models to validate data impact with an accent on rapid empirical validation and client-grounded scientific reasoning. Focus on shaping project specifications, iterating on training pipelines, and collaborating on publications and benchmarks.

Location: Hybrid model with 2 days per week in office in San Francisco Bay Area or Wrocław, Poland

Salary: $140,000–$200,000 USD (United States-based candidates)

Company

Building critical infrastructure for AI development with enterprise annotation tools, frontier data labeling services, and expert marketplace.

What you will do

• Engage as technical peer in client scoping meetings with frontier AI labs, challenging data assumptions and shaping project specs.
• Develop deep understanding of client architectures and training methods to reason about data strategies and risks.
• Run ablation studies and fine-tune open-weight models on client data to measure impact empirically.
• Consult on annotation schemas, task designs, and quality systems to ensure technical soundness.
• Collaborate with Applied Research on publications, benchmarks, and generalizable findings from client work.

Requirements

• MS or PhD in Machine Learning, NLP, Computer Science, or related quantitative field.
• Hands-on experience fine-tuning large language models (Llama, Mistral, Qwen, etc.).
• Strong understanding of LLM training pipelines (pretraining, SFT, RLHF/DPO) and data impact.
• Experience designing and executing rigorous experiments with hypothesis testing and statistical analysis.
• Ability to operate at high speed from problem to results in days.
• Strong written and verbal communication for client presentations and publications.

Nice to have

• Prior experience at frontier AI lab or applied ML startup with client interaction.
• LLM evaluation and benchmarking (metrics, eval harnesses).
• Familiarity with human data pipelines and quality assurance.
• Experience with RL, reward modeling, or RLHF.
• Published research in ML/NLP.

Culture & Benefits

• High-impact, early-stage startup environment with focus on ownership and rapid execution.
• Technical excellence at AI frontier, surrounded by curious minds solving complex problems.
• Hybrid work with collaboration in SF or Wrocław offices.
• Continuous growth tied to contributions, with 25–30% time for research collaboration.
• Autonomy, clear ownership, and metrics-driven results.