Machine Learning Engineer

Latent

San Francisco, California, United States

San Francisco, California, United States

Apply Now

About

Latent is building the intelligence infrastructure for American healthcare. Our products are already helping hospitals and clinics dramatically increase workflow output, speed up patient access to medications, and boost provider revenue. Our flagship multi-modal search and question-answering platform analyzes EHR data to surface the most relevant information, reducing operational overhead and improving care delivery. We’re a small, mission-driven team backed by General Catalyst, Conviction, and YC, tackling some of healthcare’s hardest technical challenges. If you're passionate about applying cutting-edge machine learning in a high-stakes domain, we’d love to meet you. About the Role As a Machine Learning Engineer at Latent, you’ll design and deploy advanced models at the frontier of medical language understanding. You will develop systems that can interpret long-form clinical text, generate auditable justifications for medical decisions, and reason over structured and unstructured data to automate the prior authorization process end-to-end. You’ll work on some of the most pressing problems in applied AI—balancing model expressiveness with verifiability, maintaining safety in open-ended generation, and scaling LLMs to production in high-stakes environments. This is a rare opportunity to bring research into production at the edge of what’s possible in medicine and AI. This is a high-impact, high-ownership role based full-time onsite in our San Francisco office. What You’ll Do Train and fine-tune large open-source language models for clinical reasoning, medical question answering, and evidence-grounded generation, where the stakes are human health
Design and scale multimodal embeddings to encode clinical documents, structured EHRs, and payer policies in a unified space
Own the lifecycle of ML systems—from research prototypes to fault-tolerant, privacy-compliant services running in production
Build robust retrieval pipelines for real-time semantic search and RAG architectures in the clinical domain
Collaborate with clinicians, engineers, and product leaders to ensure outputs are interpretable, auditable, and aligned with real-world constraints
Contribute to a culture of ML excellence through code reviews, experimentation frameworks, and internal knowledge sharing
You Might Be a Fit If You… Have a strong foundation in ML research or systems—through an advanced degree or high-impact work in industry
Have deep experience with NLP and LLMs (e.g., fine-tuning, LoRA, RAG, quantization) and are comfortable with frameworks like PyTorch, Hugging Face, and LangChain
Have shipped ML models in production environments—especially where latency, safety, or interpretability were critical
Are excited by ambiguous, zero-to-one problems and can think creatively about tradeoffs between performance, explainability, and reliability
Thrive in fast-moving, ambiguous, enjoy working on open-ended technical challenges, and work well with minimal oversight
(Bonus) Have published ML research or contributed to the broader ML community
(Bonus) Have worked with clinical, biomedical, or claims data—or are excited to learn the domain deeply
Why You Should Join Us Backed by top investors: General Catalyst, Conviction, and YC
Tight-knit, world-class team with a deep sense of mission
Huge greenfield opportunity with significant ownership and room for growth
Competitive salary and equity compensation. The equity upside of an early-stage startup with the product-market fit of a later-stage company.
Excellent benefits and versatile health, dental, and vision coverage plans
Paid parental leave
Lunch and dinner provided at the office
Unlimited PTO
#J-18808-Ljbffr

San Francisco, California, United States

Languages

English

Notice for Users

This job comes from a TieTalent partner platform. Click "Apply Now" to submit your application directly on their site.

Apply Now