About Playbook

Playbook is building the Operating Memory for regulated industries.

In life sciences and other highly regulated sectors, companies run on thousands of written procedures, SOPs, work instructions, policies, manuals, and how-to guides. Employees are expected to find, understand, follow, and stay trained on the right procedure at the right moment. That system is breaking under the weight of information overload.

Playbook turns procedural documents into structured, usable data making critical operational knowledge consumable by humans, software systems, and AI agents. This living layer of procedural knowledge is what we call Operating Memory.

We work with leading global life sciences companies and are building the foundational data model for procedural knowledge in regulated industries. We’re a Series A founder-led startup with high ambition, high standards, and a strong bias toward execution.

Why now

Enterprise AI agents can’t run on fragile data pipelines. They need deterministic, auditable, and low-latency infrastructure under the hood.

Your Role

Our platform runs on the Playbook® Operating Memory, an ontology-based knowledge layer that turns company rules, procedures, and regulatory requirements into explicit, machine-executable logic. AI runs through every step of it, from reading and parsing documents, to building and linking the knowledge layer, to answering questions and automating workflows. Every output can be traced back to its source.

We are hiring a Senior AI Engineer to own the AI systems across that whole path. You don’t need to be a knowledge graph specialist, your job is to work across the system from ingestion to the final answer, build it from clear data models and modular components, test ideas as hypotheses against real data, turn what works into product that customers rely on, and use LLMs only where they add leverage, never as the source of truth.

You will report directly to our Head of AI and operate as a senior Individual Contributor, collaborating closely with our cross functional teams. Success looks like AI components that can each be measured, fixed, or replaced, quality that is measured rather than assumed, and a product that teams trust without babysitting it.

What you’ll own

• Design and optimize the AI pipelines across the platform (document parsing and extraction, retrieval, vector stores, knowledge graphs, multi-agent workflows) to index and query complex operational documentation and enterprise SOPs

• Fine-tune and orchestrate LLMs to reliably translate natural language inputs into deterministic, structured API calls and workflow automations

• Build rigorous evaluation frameworks to measure and reduce hallucinations, enforce schema validation, and ensure factual accuracy for business processes, with a baseline in place before any optimization

• Deploy, scale, and monitor production AI microservices on cloud infrastructure (high availability, low latency, cost efficiency)

• Ensure all AI models and data pipelines comply with GDPR, strict enterprise access controls, zero-data-retention rules, and complete trace auditability

• Own the technical design and delivery of key parts of the AI system (with clear constraints, but real decision authority), and contribute to team-wide design decisions and code review

What you’ll bring

• 5+ years building production systems involving ML, search, or LLMs, including LLM-based systems you shipped and evaluated

• Strong CS fundamentals (data structures, algorithms, system design)

• Experience in regulated or high-stakes environments (healthcare, pharma, finance, compliance)

• Exposure to: knowledge graphs, ontologies, linked data, search engines, information extraction

• Comfort with greenfield systems and long-term ownership

• Fluency with coding agents and modern AI tooling in everyday engineering (we build with the same technology we ship)

Nice to have

Tech environment (indicative, not dogmatic):

• Python

• LLMs with a focus on reliability and control

• Search, vector stores, and graph databases

• AWS

Probably Not a Fit

This role likely isn’t for you if you…

• You primarily want to experiment rather than ship

• You optimize for personal comfort, not system impact

• You need heavy process, detailed roadmaps, or constant alignment

• You avoid conflict or direct feedback

• You trust a good demo more than a measurement