Founding Machine Learning Engineer

• →Building a production pipeline of multi-step, tool-using visual reasoning agents that run on smart glasses against real industrial workflows such as inspections and standard operating procedures
• →Creating real-time voice and video AI interfaces for the glasses, including conversational and proactive alert modes tailored to different users
• →Owning evaluation and the data flywheel: eval harnesses, capturing failure modes, and turning customer data into fine-tuning loops that improve model quality release over release
• →Delivering edge inference and model orchestration that adapts gracefully to changing connectivity and latency constraints in the field
• →Fine-tuning and optimizing open-source multimodal models (SFT, RLHF, quantization) for on-premise enterprise deployments

Tech stack: Python, PyTorch, vLLM, Triton, Ray Serve, ONNX, TensorRT, Hugging Face Transformers, LangChain, RAG, RLHF, SFT, quantization (GPTQ, AWQ), edge AI, multimodal LLMs, vision-language models, Docker

• Up to 3 years of hands-on experience building and deploying multimodal or vision-language AI systems in Python and PyTorch
• A track record of shipping vision-language systems that real users rely on in production, where you owned both the model and the orchestration layer, rather than demos or research prototypes
• Practical depth in applied model work such as fine-tuning (SFT, RLHF), evaluation design, and orchestrating models in production, including visual reasoning and detection or segmentation where needed
• Experience building agents that plan across multiple steps and call tools, along with the evaluation harnesses and data loops that keep them improving
• Experience as a founder or very early engineer at a startup, or at a fast-paced, high-intensity engineering organization
• A bachelor's or master's degree in computer science, machine learning, or engineering from a strong program, or equivalent experience shown through production AI work
• Real enthusiasm for computer vision, wearables, and industrial AI, visible in your projects, side work, or career path
• Ability to work on-site in San Francisco five days a week; openness to shared team housing is a plus
• Existing US work authorization; our client can support visa transfers (for example OPT or H-1B transfer) but cannot sponsor new visas

• Experience shipping AI for AR or wearable devices, or computer vision for autonomous driving
• A master's degree that included vision or multimodal research, such as a thesis or published work
• Experience deploying open-weight models at the edge or on-premise on constrained hardware using tools like vLLM, Triton, TensorRT, or quantization techniques

• Location: San Francisco, CA
• Work policy: On-site 5 days per week
• Compensation: $180,000–$230,000 + equity
• Visa sponsorship: No new sponsorship; open to visa transfers (for example OPT or H-1B transfer)
• Employment type: Full-time