Founding AI Engineer

About Hyperspell

Hyperspell is the Memory & Context Layer for AI Agents. AI agents are clueless geniuses. AI agents are crushing humans on any standardized test, but wouldn’t last a day at a real job. What today’s super‑intelligent agents are missing is the real world context they are operating in. A context that humans stitch together from hundreds of data points across dozens of interactions and channels.

A context that grows with their tasks. Hyperspell gives AI agents this context by connecting to their user’s workspace data and building a personalized memory and context layer.

We believe the future is shaped by those with the boldest ideas. Our mission is to become the default context layer for all organizations that all their AI agents can connect to. We’re not doing this because it’s easy, but because it’s worth it. We care deeply. We care about our customers and their success, about the future of AI and its contribution to society, and most of all, about each other.

Building the frontier of AI, we are de‑facto one of the most technologically advanced companies in human history — and because of that, not despite that, want to be the most human company too. A place where we can enjoy shaping the future together.

As our founding ML engineer, your mission is to advance what’s possible with AI agents. You’ll be working at the frontier of how AI systems understand, remember, and reason about information.

• Contextual knowledge: building systems that extract entities and relationships from unstructured data, enabling agents to traverse connections, surface insights, and provide context across organizations.
• Smarter retrieval: developing hybrid search strategies that combine semantic similarity, graph traversal, and other heuristics.
• Memory that evolves: creating systems that don’t just store information but understand when facts change, determine what’s relevant, and intelligently forget what’s no longer valid.
• Query understanding: teaching our system to expand and rewrite queries, understand user intent, and retrieve information the user needs but didn’t know to ask for.
• Multi‑modal pipelines: improving how we parse, chunk, and represent complex documents, structured and unstructured data, tables, images, call recordings, hierarchical content.

This is a high‑autonomy role where you’ll own problems end‑to‑end: from researching approaches, to prototyping, to shipping production systems that handle millions of documents. You’ll work directly with founders and have outsized impact on the product and company direction. You don’t need experience training models from scratch, but you need a deep understanding of LLMs, embeddings, NLP, and agentic architectures; as well as an innate intuition of what actually moves the needle on agentic quality.

• Within 30 days, you’ll own our core agentic query loop — the experience of serving our customer’s AI agents the right context at the right time.
• Within 60 days, you’ll work on our most ambitious projects: building a continuously updating, self‑referential memory network; building procedural memory and memory for multi‑user agents.
• Within 90 days, you’ll product ionize our memory features, present your work at conferences, and will already be working on the problems AI agents will be facing a year from now.

• 3‑5 years of experience in applied machine learning, AI engineering, or natural language processing.
• Fluency in Python.
• Clarity of thought, excellent communication.
• Ability to work on location in San Francisco.

• $150-250k base salary.
• 0.5-1.5% equity.
• $10,000 to build your dream work setup.
• Health, dental, and vision coverage and generous health benefits.