Job Description, Responsibilities & Requirements
About the Position
We are seeking a Lead AI Engineer to serve as the technical owner of our connector implementation, responsible for designing and building the Drug/Drug Interaction Smart Connector in vanilla LangChain and LangGraph with no platform-specific dependencies. In this role, you will architect agentic orchestration that transforms natural-language questions into structured queries, lead a proof-of-concept validation and collaborate with cross-functional experts to ensure statistical and clinical correctness.
Responsibilities
- Design and build the Drug/Drug Interaction Smart Connector in vanilla LangChain and LangGraph with no platform-specific dependencies
- Architect the agentic orchestration that turns natural-language questions into structured queries against the PharmaPendium Advanced Search API
- Ensure retrieval through the FAERS module endpoint is reliable and traceable
- Lead the lightweight POC that validates the implementation can be cleanly refactored into SAIP platform
- Resolve compatibility issues before they become blockers
- Build out the full connector logic including candidate drug-pair generation ranked by interaction strength
- Implement MedDRA classification across generic and clinical phenotype terms
- Compute PRR/ROR disproportionality producing consistent traceable output tables with case counts, statistical measures and the signal-only interpretation flag
- Enforce clean well-documented code structured to support Codex-assisted refactoring
- Partner closely with the Data Scientist and Pharma SME to ensure statistical and clinical correctness
- Tune the connector against the agreed Golden Q&A set so outputs meet the acceptance baseline through validation and handover
Requirements
- 5+ years of experience in AI engineering or related technical roles
- Expertise in vanilla LangChain (v1.2.6) and LangGraph (v1.0.7)
- Proficiency in designing agentic orchestration that converts natural-language questions into structured queries
- Skills in integrating with the PharmaPendium Advanced Search API and FAERS module endpoint
- Knowledge of MedDRA classification across generic and clinical phenotype terms
- Competency in PRR/ROR disproportionality computation and statistical measures
- Capability to lead POCs and resolve compatibility issues for platform refactoring
- Background in writing clean well-documented code structured to support Codex-assisted refactoring
- Understanding of collaborating with Data Scientists and Pharma SMEs to ensure statistical and clinical correctness
- English proficiency at B2 level or higher
We Offer
- Remote work in Türkiye