Forward Deployed Software Engineer

KEY RESPONSIBILITIES

Embedded Client Delivery

• Embed directly within pharma client organisations operating as a trusted technical peer, not a vendor — and own the design, build, and deployment of production AI systems end-to-end on the client's own infrastructure
• Lead the technical workstream and direct teams of engineers — Agent Pipeline Engineers and AI-Augmented Engineers — under your architecture and delivery ownership
• Hold the technical client relationship at Director, VP, and CDO level: scoping problems, presenting architecture trade-offs, defending design decisions under scrutiny, and translating technical outcomes into business language

AI Architecture and Engineering

• Architect multi-agent AI systems for pharma environments — spanning orchestration patterns, tool and function integration, retrieval-augmented generation, memory architectures, human-in-the-loop design, evaluation pipelines, and production MLOps
• Build on the technology stacks clients already operate: Databricks (Delta Lake, Mosaic AI, Genie), Snowflake (Snowpark, Cortex AI, Cortex Analyst), AWS (Bedrock Agents, SageMaker), and the Claude and Anthropic API stack with Model Context Protocol
• Design and implement AI evaluation frameworks appropriate for regulated pharma environments — probabilistic quality thresholds, RAGAS, LLM-as-judge, adversarial red-teaming, and audit-trail-compliant output governance
• Ensure systems are production-grade: observable, maintainable, secure, and compliant with pharma data governance requirements including HIPAA, GDPR, and applicable FDA AI/ML guidance
• Stay current with the agentic AI ecosystem — frameworks, model capabilities, evaluation techniques, and orchestration protocols — and translate emerging capability into deployment-relevant technical decisions

Pharma Domain Translation

• Translate pharma commercial and clinical business problems into AI-solvable architectures without requiring a domain primer from the client — the depth of your domain expertise is part of what you bring to the engagement
• Validate that AI outputs are accurate against pharma business logic and commercial or clinical norms — not just technically correct but domain-defensible and explainable to the end users who act on them
• Serve as the connective layer between the client's business problem and the technical solution, eliminating the scoping ambiguity that causes most pharma AI deployments to stall before production