AI Engineer - Relational Foundation Models & Agentic Systems

Why this role (and why now)

With the launch of our new Kumo Relational Foundation Model (RFM), we’ve seen unprecedented interest from builders who want to create on top of our platform. This is your chance to be part of that momentum. You’ll be building applications and agentic workflows, demoing them to customers, and making core product and engineering decisions that shape the growth of our RFM offering.

This is an engineering role at its core, but you’ll also interface directly with customers, the broader builder community, and cross-functional engineering teams. We’re looking for someone who thrives in that hybrid space—shipping product, representing engineering externally, and helping shape the direction of Kumo RFM from the ground up.

We’re looking for someone who’s:

• Comfortable in an innovation pod or startup environment, moving quickly from idea → prototype → ship.

• A tinkerer at heart who’s built full‑stack apps (frontend, backend, data) and lately has been hands‑on with the LLM tooling ecosystem.

• Collaborative and easy to work with—you know how to partner with PMs/design/ML, bounce ideas, and get things done together.

Bonus: experience as a Founding Engineer or early builder who has shaped product direction from the ground up.

About Kumo.ai

Kumo.ai is redefining enterprise AI with foundation models for relational data, enabling organizations to predict, optimize, and act with speed and confidence. Our agentic systems collaborate with data teams, turning complex business tables and SQL workflows into interpretable, actionable, and automated insights.

The Role

We’re hiring an AI/ML Engineer to design and build AI-powered, user-facing products on top of our RFM. You’ll ship features that:

• Understand a user’s goal and autonomously propose workflows for analysis, prediction, and optimization.

Interact with enterprise systems and APIs, orchestrating tools and data.

• Produce interpretable outputs that are easy to trust in real-world decisions.

• Demo prototypes and apps to customers, iterating on feedback and incorporating real use cases.

You’ll work across product surfaces (UI/API), agent orchestration, data/infra, and model integration, collaborating tightly with product, design, ML research, and customers.

• Design, implement, and deploy AI agents that assist data scientists on relational/SQL data and recommend next-best actions.
• Build user-centric APIs and product surfaces (web/UI or programmatic) that make agentic workflows feel seamless and reliable.
• Integrate Kumo’s Relational Foundation Model with enterprise data systems; contribute to tooling, retrieval, and guardrails.
• Develop adaptive, multi-step workflows (LLM orchestration, tool use, feedback loops) that continuously refine outputs.
• Ensure interpretability and evaluation: traceability of steps, confidence scoring, and human-in-the-loop review.
• Collaborate with PM/design/ML research to turn ambiguous problems into shippable product; instrument, measure, iterate.
• Demo your work to customers and community, serving as a visible builder and advocate for Kumo RFM.
• Optimize for latency, cost, and reliability in production environments (serving, caching, tracing, observability).

• 1+ years in ML/AI product development or software engineering (startup or fast-paced product teams).
• Hands-on with embeddings, vector databases, and RAG; practical experience evaluating retrieval quality.
• Strong background in deep learning/transformers/foundation models and LLM orchestration (tool use, planning, memory).
• Experience with relational data & SQL; structured reasoning on business datasets.
• Proficiency in Python and familiarity with data wrangling (Pandas, NumPy).
• Strong product sense and collaboration skills—comfortable working with PMs/design and iterating with users.

• Experience as a Founding Engineer or early builder at a startup/innovation pod.
• Experience with LangChain, LangGraph, LlamaIndex, OpenAI/Anthropic APIs, and multi-agent coordination libraries.
• Track record building full-stack features (you can dip into frontend/backend/data/infra as needed).
• Experience integrating agents with enterprise systems and APIs; designing foundation APIs for tools.
• Background in knowledge graphs, GNNs, causal inference, or structured reasoning with LLMs.
• MLOps and cloud (AWS/GCP), model/agent serving, prompt/runtime observability, and eval pipelines.
• Familiarity with guardrails, safety, and governance for enterprise AI.

• Builder mindset: you ship small, learn fast, and raise the bar with each iteration.
• Team-first: crisp communication, lightweight docs, and pairing with PM/design/ML to de-risk quickly.
• User-obsessed: you instrument everything, close the loop with customers, and let usage guide the roadmap.
• Pragmatic about research: you know when to use existing models vs. when to push the frontier.

• Ship a v1 agentic workflow powered by RFM that users adopt for a real analytics task (with instrumented evals & feedback).
• Demonstrate measurable improvements (accuracy/latency/cost/UX trust) via experiments and A/Bs.
• Land 1–2 integrations with enterprise data systems or tooling; document runbooks and guardrails.
• Deliver customer demos that inspire adoption and open new product opportunities.

• Python, Pandas/NumPy, PyTorch/JAX/TensorFlow, LangChain/LangGraph/LlamaIndex, OpenAI/Anthropic APIs, vector DBs, SQL, modern web stack (we’ll meet you where you are), AWS/GCP, observability/tooling for agents.

Why join Kumo

• Shape the future of AI for structured data with a truly differentiated Relational Foundation Model.

• Build net-new agent experiences that make data teams faster and more effective.

• Work with a small, senior team that ships—high impact, low bureaucracy.

• Be an early voice in growing Kumo’s RFM product—influence direction, ship demos, and help define the category.

Kumo.ai is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.