Head of AI Research

Introducing Kepler

High-stakes industries are falling behind on AI adoption. Their workflows can’t afford wrong answers, and AI can’t be trusted to give right ones because of hallucinations. The barrier isn’t that the models aren’t smart enough; it’s that no one can verify what they produce. The fix isn’t a better model, it’s a trust layer: every output traceable, every calculation auditable, every answer reproducible. Kepler is the agent harness – the infrastructure layer that wraps around AI models to make their outputs reliable, traceable, and verifiable. The model is a replaceable component; the harness is the product.

In Kepler's architecture, the LLM orchestrates – it decides what data to gather, what to compute, how to structure the output. But every actual data point, every extracted value, every calculation flows through deterministic code pipelines. The LLM never touches the data itself. Every value carries provenance metadata back to its exact source. Every computation is auditable and reproducible. Verification loops cross-check outputs before users ever see them.

We started in finance because the stakes are highest and the tolerance for error is zero. We’ve built a finance research product that lets analysts supercharge their workflow: pulling comparables, building models, and researching filings. No more double-checking every number AI spits out. Every number tracing back to the source, every time.

But the architecture – provenance, deterministic computation, verification – applies anywhere trust in AI output matters: chemicals, legal, healthcare. Models are commoditizing fast. The trust layer is what's missing, and the market is massive.

The Role

What You'll Own

You'll define the research agenda that makes AI trustworthy for enterprise decisions. At Kepler, we've solved hallucination not by making models smarter, but by architecting systems where hallucination is structurally impossible. AI interprets intent. Code retrieves data. A semantic layer connects them. Every output traces to its source.

Now we need someone to push the frontier of what's possible within this architecture.

You'll lead research across agentic systems, memory architectures, retrieval mechanisms, and evaluation frameworks. You'll have access to completely differentiated financial data: structured filings, earnings transcripts, market feeds, research reports, live audio, all normalized with full provenance. This isn't another research lab building demos. Your work ships to production, powering research workflows where financial professionals make million-dollar decisions.

This role is for researchers who want to build AI systems that actually work in high-stakes environments, where every answer must be defensible and every insight must trace back to truth.

Key Responsibilities

• Define the research agenda: Identify the highest-leverage research problems at the intersection of AI reliability, agentic systems, and enterprise trust. Shape what Kepler builds next.
• Pioneer agentic architectures: Design systems for agent orchestration, task planning, and multi-step reasoning that maintain accuracy and provenance across complex research workflows.
• Advance memory and retrieval: Build novel approaches to context management, knowledge representation, and retrieval that let agents reason over millions of documents while staying grounded in sources.
• Build evaluation frameworks: Define how we measure agent quality, reliability, and trustworthiness. Create benchmarks that matter for real-world financial research.
• Ship research to production: This isn't theoretical. Your work deploys to systems used by top financial institutions. Research validates through real-world impact.
• Build and lead the research team: Hire exceptional researchers. Create an environment where breakthrough work happens and ships.
• Publish and represent Kepler: Contribute to the broader AI community through publications, open-source work, and presence at top venues like NeurIPS, ICML, and ICLR.
• Collaborate with founders: Work directly with leadership on technical strategy, product direction, and company vision.

What We're Looking For

• PhD in Computer Science, AI/ML, or related field
• 5+ years in AI/ML research with focus on LLMs, agents, retrieval systems, or memory architectures
• Publication record: Strong track record at top venues (NeurIPS, ICML, ICLR, ACL)
• Production experience: You've shipped research to real systems with real users. You understand the gap between paper and production.
• Technical depth: Expertise in transformer architectures, attention mechanisms, retrieval-augmented generation, or memory-augmented systems
• Leadership: Experience building and leading research teams in high-stakes environments
• Research taste: Ability to identify problems that are both technically interesting and practically important
• Strong communicator who can translate complex research into clear implications for product and business
• Intellectually humble. Confident enough to lead in your domain, wise enough to collaborate across disciplines.
• Energized by hard problems where the answer isn't known

Why This Role is Different

• Differentiated data: Access to structured financial data with full provenance that doesn't exist anywhere else. Train and evaluate on real, high-quality data.
• Architecture that matters: We've built the infrastructure that makes AI trustworthy. Your research extends what's possible within a system designed for accuracy.
• Research that ships: Your work powers decisions at top financial institutions. Impact is measured in production, not citations.
• Greenfield opportunity: Define the research agenda for a company building the ground truth platform for AI.
• Founder-level collaboration: Work directly with founders who built Palantir Foundry and data infrastructure at Citadel.

Research Benefits

• Open publication policy: We actively encourage publishing at top venues and contributing to the broader AI community. Your research belongs to the field.
• Dedicated GPU capacity: Pre-reserved compute for research experiments. The computational resources to pursue ambitious ideas.