AI Research Engineer

Edge Tech•San Francisco, CA

7d•$250,000•Hybrid

About The Position

This isn’t your typical research gig. We’re working with a stealthy, fast-growing AI startup that’s redefining how enterprises understand and act on complex documents. Their platform powers 200M+ documents per month , drives 4M+ open-source downloads , and is trusted by Fortune 500 clients across industries. They’re looking for an AI Research Engineer (3–4 yrs experience) to join the core document understanding team, someone smart, adaptive, and hungry to push the state-of-the-art in document AI. Bonus if you’ve worked on vision-language models, document processing, or agentic AI systems.

Requirements

Strong ML Foundation: 3–7 yrs in ML engineering or applied research, with experience benchmarking and training models.
Document AI / Vision Models: Hands-on experience with document understanding, vision-language, or agentic systems.
Python & ML Tools: Proficiency with PyTorch and modern ML frameworks.
Early-Stage / Research Pedigree: Startup, academic, or research lab experience (ex-founders welcome).
Soft Skills: Scrappy, adaptable, and able to communicate complex technical ideas clearly.

Nice To Haves

vision-language models
document processing
agentic AI systems

Responsibilities

Build & Train Cutting-Edge Models: Vision-language models for document processing, data curation, synthetic data generation, and benchmark creation.
Post-Training Excellence: Evaluate and fine-tune base models to meet strict performance and cost benchmarks.
Agent Innovation: Iterate on the agent layer atop base models to improve efficiency, accuracy, and usability.
Customer Insight: Collaborate with clients (<15% of your time) to translate real-world needs into model capabilities.
Experiment & Innovate: Push boundaries of ML research, improve experimentation velocity, and develop novel methods that beat the status quo.