Head of AI

Highlight AI•San Francisco, CA

22h•$250,000 - $350,000•Onsite

About The Position

Highlight is building a shared intelligence layer for the modern workforce. Highlight unifies context across every person and tool on your team, seamlessly bridging information silos. It evolves with your organization to proactively route knowledge and reliably automate workflows. The Role We're hiring a Head of AI Engineering to lead and scale our ML and AI efforts as we build the engine that powers Highlight. You'll own the architecture, quality, and roadmap for how we ingest, model, and surface context, starting from data pipelines and evals to the context graph that makes our system truly useful. This is a hands-on leadership role: you'll write code, set technical direction, mentor engineers, and collaborate closely with Backend and Product teams to ship features that feel like magic. We're looking for a technical leader who's equally comfortable diving into model architectures and rallying a team around ambitious goals. You have deep expertise in LLMs, knowledge systems, and retrieval. You're energized by the challenge of building something new. You thrive in ambiguity, ship fast, and hold yourself and your team to a high bar. Note: this is an on-site role, requiring five days a week in our office.

Requirements

7+ years in ML/AI engineering, with 3+ years leading teams or major technical initiatives
Master’s or PhD in Computer Science, Machine Learning, Statistics, Applied Math or equivalent experience.
Published research (or experience shipping to production) in knowledge graphs, memory systems, retrieval, or continual learning
Experience working at top AI labs (e.g. DeepMind, OpenAI, Anthropic, Meta FAIR, xAI) or leading AI application companies (e.g. Google, Cursor, Perplexity)
Strong architectural chops: you've designed and scaled ML infrastructure from scratch
Hands-on proficiency with modern ML stacks (PyTorch, TensorFlow, RAG, vector DBs, cloud infra)
A player-coach mentality: you love mentoring engineers and getting your hands dirty in the code
Ownership mentality (no task is too small for you) – it’s all hands on deck here and we need everyone rowing the boat.
Excellent communication skills and a collaborative mindset – we’re extremely lean so we’re looking for a team player who works well with others.
Type-A tendencies: maybe you’re extremely detail-oriented. Or maybe you take pride in shipping high-quality products. Whatever it is you do, you give it your all.
Must be based in or willing to relocate to SF – although we’re flexible with days off & schedules, we have a 100% in-office culture during the week.
Hustle is expected, grit is required.

Nice To Haves

We're looking for someone who is interested in building the AI user experience for the future.
We punch above our weight so we value extreme ownership, accountability, and proactiveness.
The ideal candidate has deep, experienced understanding of LLMs or Knowledge systems, and a track record of shipping production AI systems at scale.
And even if you don’t fall neatly into any of these buckets, we’re looking for scrappy and motivated self-starters above all.

Responsibilities

Own the context graph: architect and build Highlight's memory systems end-to-end, including data ingestion, modeling, retrieval, and evals
Research, prototype and deploy state-of-the-art machine learning text and OCR models (e.g., transformer architectures, computer vision, time series forecasting, reinforcement learning)
Lead the full ML project lifecycle: data pipelines, training, evals, deployment, monitoring, scalability, latency, retraining, A/B testing, and model lifecycle management
Own output quality across Highlight chat, actions, and workflows; drive toward industry leading acceptance rates.
Build and lead a world-class ML Engineering team; Establish AI/ML best practices across the broader engineering org.
Stay current on knowledge representation, continual learning, and retrieval; translate insights into features that feel like magic; Research novel approaches and publish when appropriate.