14 days ago
Base Salary
$175k - $250k/yr
Responsibilities
- Train and fine-tune models for memory extraction, updates, consolidation, forgetting, and conflict resolution.
- Convert research literature into prototypes, benchmark approaches against baselines, and productionize the strongest results.
- Build automated relevance, accuracy, and consistency metrics, gold sets, online A/B and interleaving tests, and dashboards.
- Identify customer problems, formulate testable research hypotheses, and validate them through field trials.
- Design APIs and data contracts, plan safe rollouts, and ship systems meeting latency, reliability, and cost goals.
- Orchestrate batch and streaming data pipelines for low-latency production model execution.
Requirements
- Experience building RAG or information-retrieval systems for real products, including retrieval, ranking, or query understanding.
- Experience training or fine-tuning LLMs or encoders, with strong experimental design and rapid iteration skills.
- Strong Python skills and deep PyTorch experience, plus familiarity with vLLM and modern serving frameworks.
- Experience building evaluation systems for complex language, retrieval, or generation tasks, including gold sets, offline metrics, and online tests.
- Ability to run production model data pipelines with low-latency service-level objectives across batch and streaming workloads.
- Clear communication with engineering, product, go-to-market, and customer stakeholders.
- Preferred: publications at venues such as NeurIPS, ICML, or ACL.
- Preferred: experience with privacy-preserving machine learning, including redaction, differential privacy, or data governance.
- Preferred: familiarity with memory and retrieval literature or prior memory-system work.
- Preferred: expertise with embeddings, vector-database internals, deduplication, or contradiction detection.
Benefits
- Fully covered health, dental, and vision plans for employees, with subsidized dependent coverage.
- Daily catered lunch and dinner in the office.
- Flexible paid time off.
- Equity in the early-stage company.
- Regular team happy hours and events.
- Top-tier laptop and work equipment.
- Office-first, in-person work in San Francisco.
Categories
AI ResearchML Engineering