Job Description

About Merakee BI
We are a Canadian pharmacovigilance consultancy based in Mississauga, Ontario. Pharmacovigilance is the only thing we do — safety systems, analytics and dashboards, data migration and extraction, and AI development for drug safety. Our clients are pharmaceutical and biotech safety groups who run regulated systems where every change has to be assessed, approved and evidenced. We are a small team, which means your work ships and you see who it affects.
About the Role
You will own the AI and machine learning side of our delivery work: LLM and agentic systems that read regulatory source documents, summarise and code safety cases, generate report sections, and surface signals in safety data. Twelve months in, you should have production systems running against real client data, with the pipelines, evaluation and guardrails that let a safety team trust the output — and you should be the person we point to when a client asks how the AI actually works.

What You’ll Do
• Design and ship LLM and agentic systems end to end — retrieval-augmented pipelines over regulatory documents, structured extraction, and automated narrative and report-section generation.
• Build the data layer underneath them: SQL and pipeline work over safety and clinical datasets, with the transformation and quality checks that make model output defensible.
• Apply machine learning where it beats prompting — classification, coding, anomaly and signal detection over safety data.
• Own evaluation and guardrails: define how we measure whether a model is right, and build the content-safety and review checkpoints that catch it when it is not.
• Build the application around the model — APIs, services and enough front-end to put a working tool in a reviewer’s hands.
• Work directly with our PV consultants and client safety teams to turn domain requirements into systems, and explain technical trade-offs to people who are not engineers.

Required languages: English

Education level: Bachelor degree in statistics/Software engineering/Mathematics or related fields

Required skills: • 7+ years building and shipping production software, with recent focus on machine learning and AI. • Hands-on experience with production LLM systems — RAG, retrieval and chunking strategies, structured output, agent or tool-calling patterns, and prompt evaluation. • Strong Python. • Strong SQL, and experience building data pipelines over substantial datasets. • Full-stack delivery experience — REST APIs and a modern JavaScript framework — enough to build and deploy a working application, not only a model. • Experience taking something from prototype to production: deployment, monitoring, and dealing with it when it breaks.

Closest intersection: Mavis/Britania