← Search all jobs

Founding ML Research Engineer

HSF.io

San Francisco, CA · Posted November 13, 2026

Just laid off? Start here: a plan for the first weeks

See every open role at HSF.io

More in California job market · Technology & Software · Data & Analytics jobs

More roles at HSF.io

Similar jobs

Where San Francisco, CA ranks

Where San Francisco, CA placed in our monthly rankings, month by month. Each month stays online at its own address.

How AI is used in this kind of work

Among computer and information research scientists, AI is already used for a large share of the tasks. Measured exposure for computer and information research scientists: 34% (high).

Share of this job’s tasks that AI is already used for at work, from real usage. It is not a forecast of job loss, and a low number means no use seen yet. Source: Anthropic Economic Index.

Check how this applies to your own job

About the job

Seed · ~5 people · $5M raised · Wearable AI co-pilots on smart glasses for industrial field workers, deploying agentic vision-language models to…

San Francisco · In-office, full-time

About the company

The company is building wearable AI co-pilots for industrial field workers, deploying smart glasses powered by agentic vision-language models (VLMs) to legacy industries like data centers, energy grids, and aerospace. The company is on a mission to bring AI into environments where technology—let alone artificial intelligence—has never meaningfully helped before.

The company combines deep expertise in mechanical engineering, AI, and robotics with a long-term vision for industrial automation. The company operates at the intersection of computer vision, multimodal AI, and wearable hardware, collecting real-world data that can serve robotics automation down the line. The founders are joined by exceptional early hires: Stanford PhD dropout and olympiad medalist & ex-engineer.

Culture: the company runs with the intensity of a top-tier style startup. The team values hunger, scrappiness, and genuine mission alignment over pedigree. Engineers are expected to operate with high ownership, move fast, and iterate constantly as the AI landscape evolves. The founders prioritize people who are excited about computer vision and industrial applications—not just chasing the latest LLM trends. The work environment is collaborative, in-person, and high-tempo, with a strong emphasis on shipping production systems to real customers.

With anchor accounts at world-leading data center and energy grid companies already on path to $10M ARR and 10-15 enterprise prospects in pipeline (including hyperscalers, top aerospace primes, and US Army), the company is solving one of the most interesting applied AI problems in the market today.

The role: Founding ML Research Engineer

We're deploying wearable AI co-pilots to data centers, energy grids, and aerospace — as a foundational member of the engineering team. This is a rare opportunity to own the full multimodal AI stack from day one: production VLM pipelines, eval frameworks, edge inference, and real-time voice-video AI running on actual hardware in the field. You should be a scrappy, high-ownership builder who has shipped agentic or multimodal systems to real users, not just demos. If you're excited about computer vision, wearables, and solving hard applied AI problems in legacy industries, this role is for you.

Backing and traction: $5M seed from leading investors. Two anchor customers (world-leading data center and energy grid companies) are already on a path to $10M ARR, with 10–15 more enterprise prospects in the pipeline, including hyperscalers, top aerospace primes and the US Army.

What will you be doing?

Ship a production agentic-VLM pipeline on smart glasses: multi-step, tool-using visual-reasoning loops running against real industrial workflows (SOPs, inspections)

Build and own the eval + data flywheel: eval harnesses, failure-mode capture, and customer-data → fine-tune loops that measurably improve model quality each release

Ship field connectivity adaptation and edge inference: model orchestration, runtime optimization, and graceful fallback across varying connectivity and latency budgets

Build real-time voice-video AI interfaces for the glasses: video-heavy, conversational speech, and proactive alert modes adapted to different end-user profiles

Drive multimodal model training including open-source model fine-tuning (SFT, RLHF, quantization) for on-premise enterprise deployments