Digital R&D Principal Engineer

About Sanofi

Sanofi is a global biopharmaceutical company dedicated to chasing the miracles of science to improve people's lives. Our Digital R&D Software Engineering team sits at the intersection of cutting-edge technology and life-changing medicine - building the digital backbone that accelerates drug discovery, clinical development, and patient outcomes. We are guided by our values: Aim Higher, Act for Patients, Be Bold, and Lead Together.

Position Summary

We are looking for a talented and driven AI Software Engineer – Agentic Solutions to join Sanofi's Digital R&D Software Engineering team. In this mid-level role, you will design, develop, and deploy agentic AI systems that autonomously reason, plan, and act to solve complex challenges across pharmaceutical research and development.

You will work at the forefront of applied AI - building multi-agent architectures, LLM-powered workflows, and intelligent automation pipelines that directly support Sanofi's mission to bring life-saving therapies to patients faster. This includes hands-on experience with platforms such as AWS Bedrock and AWS AgentCore for scalable agent deployment, and MCP (Model Context Protocol) for standardized tool and data source integration. This is a high-impact role for an engineer who is passionate about autonomous AI systems and wants to apply them in a meaningful, regulated, and scientifically rigorous environment.

Key Responsibilities

Agentic AI Design & Development

Design, build, and deploy autonomous AI agent systems capable of multi-step reasoning, planning, and task execution across R&D workflows

Develop multi-agent architectures where specialized agents collaborate, delegate, and coordinate to solve complex pharmaceutical problems

Implement tool-using agents that interact with APIs, databases, internal systems, and external data sources

Build feedback loops and self-correction mechanisms to improve agent reliability and accuracy over time

Leverage AWS AgentCore for agent lifecycle management, memory persistence, and tool integration - deploying and managing agents at scale in a governed, enterprise environment

Design and implement MCP (Model Context Protocol)-based agent architectures, building MCP servers and clients to standardize how agents interact with external tools, data sources, and Sanofi's internal services

LLM-Based Solution Engineering

Develop and fine-tune LLM-based applications using frameworks such as LangChain, LlamaIndex, AutoGen, CrewAI, or similar

Design and implement Retrieval-Augmented Generation (RAG) pipelines to ground agents in Sanofi's proprietary scientific and operational knowledge

Engineer prompt engineering strategies, chain-of-thought reasoning, and structured output parsing for production-grade reliability

Evaluate and benchmark LLM performance across models (GPT-4, Claude, Mistral, Llama, etc.) for specific pharmaceutical use cases

Build and deploy agentic solutions using AWS Bedrock Agents and Bedrock Knowledge Bases, leveraging foundation models available on Bedrock (e.g., Claude, Titan, Llama) for scalable, managed generative AI workloads

️ Software Engineering & Architecture

Write clean, maintainable, production-ready Python code following software engineering best practices (SOLID principles, design patterns, code reviews)

Build RESTful and event-driven APIs to expose agent capabilities to downstream applications and users

Implement CI/CD pipelines, automated testing (unit, integration, regression), and monitoring for AI systems

Ensure observability of agent behavior through logging, tracing, and evaluation frameworks (e.g., LangSmith, Arize, Weights & Biases)

Systems Integration

Integrate agentic solutions with Sanofi's existing digital ecosystem including data platforms, clinical systems, ERP, and knowledge management tools

Connect agents to vector databases (Pinecone, Weaviate, pgvector, Chroma) for semantic search and memory management

Work with cloud-native services ( AWS) to deploy scalable, secure, and compliant AI workloads - including AWS Bedrock managed AI services for foundation model access and agent orchestration

Integrate MCP-compatible tools into agent workflows, enabling standardized, interoperable connections between AI models and external data sources, APIs, and enterprise systems

Collaborate with data engineers to ensure agents have access to high-quality, governed data pipelines

Collaboration & Delivery

Partner with product managers, data scientists, and domain experts (biologists, clinicians, regulatory specialists) to translate scientific needs into agentic AI solutions

Participate in Agile ceremonies (sprint planning, retrospectives, demos) and contribute to team velocity

Contribute to technical documentation, architecture decision records (ADRs), and internal knowledge sharing

Mentor junior engineers and contribute to the team's AI engineering standards and best practices

Required Qualifications

Technical Skills

Programming: Strong proficiency in Python (3.9+); familiarity with async programming, type hints, and packaging

AI/ML Frameworks: Hands-on experience with LangChain, LlamaIndex, AutoGen, CrewAI, Semantic Kernel, or equivalent agentic frameworks

LLMs & Generative AI: Practical experience building applications on top of OpenAI, Anthropic Claude, or open-source LLMs

AWS Bedrock: Hands-on experience with AWS Bedrock as a managed generative AI platform - including working with Bedrock Agents for orchestrating multi-step agentic workflows, Bedrock Knowledge Bases for RAG-based retrieval, and selecting and evaluating foundation models available on Bedrock (e.g., Anthropic Claude, Amazon Titan, Meta Llama)

AWS AgentCore: Experience using AWS AgentCore for building, deploying, and managing AI agents at scale - including agent lifecycle management, memory configuration, and tool/API integration within the AgentCore framework

MCP (Model Context Protocol): Practical experience with Anthropic's Model Context Protocol - including building MCP servers and clients, integrating MCP-compatible tools into agent workflows, and designing MCP-based agent architectures that standardize how AI models interact with external data sources and services

RAG Systems: Experience designing and implementing end-to-end RAG pipelines including chunking strategies, embedding models, and retrieval optimization

Vector Databases: Working knowledge of at least one vector store (Pinecone, Weaviate, Chroma, pgvector, Qdrant)

APIs & Integration: Experience building and consuming REST APIs; familiarity with OAuth2, API gateways, and webhook patterns

Cloud Platforms: Hands-on experience with, AWS, - including managed AI services (Azure OpenAI Service, AWS Bedrock, Vertex AI)

DevOps & MLOps: Familiarity with Docker, Kubernetes, GitHub Actions, and model deployment pipelines

Data Handling: Proficiency with SQL, JSON, and working with structured/unstructured data sources

Domain & Engineering Practices

Demonstrated experience building agentic or autonomous AI systems in a professional setting

Understanding of agent memory models (short-term, long-term, episodic), tool use, and planning strategies (ReAct, Plan-and-Execute, etc.)

Solid grasp of software engineering fundamentals: version control (Git), testing, code review, and documentation

Preferred Qualifications

Experience working in pharmaceutical, biotech, healthcare, or other regulated industries

Familiarity with GxP compliance principles and how they apply to AI/software systems

Advanced experience with AWS Bedrock - including custom model evaluation, guardrails configuration, and Bedrock Studio for rapid prototyping of agentic workflows

Experience designing MCP server implementations for proprietary enterprise data sources, enabling AI agents to securely and consistently access internal knowledge systems via the Model Context Protocol

Deeper expertise with AWS AgentCore - including advanced memory management patterns, multi-agent orchestration, and integrating AgentCore-managed agents with enterprise identity and access management (IAM) controls

Experience with knowledge graphs or ontologies in scientific domains

Exposure to fine-tuning or RLHF workflows for domain-specific LLM adaptation

Contributions to open-source AI projects or published technical writing

Experience with multi-modal AI (text + image, text + structured data)

Knowledge of clinical data standards (CDISC, HL7 FHIR) or scientific data formats

Education Requirements

Bachelor's degree in computer science, Software Engineering, Artificial Intelligence, Data Science, or a related technical field (required)

Master's degree in a relevant discipline (preferred)

Equivalent practical experience with a strong portfolio of AI engineering work will be considered in lieu of formal education

Experience Level

1–3 years of hands-on experience building AI/ML or LLM-based applications

At least 1 year of direct experience with agentic AI systems, autonomous workflows, or multi-agent frameworks

A track record of delivering production-grade AI features end-to-end, from design through deployment and monitoring

Key Competencies

Description

Technical Excellence

Writes high-quality, testable, and maintainable code; makes sound architectural decisions including selection of the right agentic platform (Bedrock, AgentCore, MCP) for the use case

Problem Solving

Breaks down ambiguous, complex problems into structured, executable solutions

Innovation Mindset

Proactively explores emerging AI techniques - including evolving agent protocols like MCP and managed platforms like AWS Bedrock - and applies them pragmatically

Collaboration

Works effectively across disciplines - engineering, science, product, and business

Communication

Clearly articulates technical concepts to both technical and non-technical stakeholders

️ Responsible AI

Considers ethical implications, bias, safety, and compliance in every design decision

Ownership

Takes end-to-end accountability for features from ideation to production

What We Offer

At Sanofi, we believe that people who chase the miracles of science deserve Extraordinary support. Joining our Digital R&D Software Engineering team means:

Global Impact - Your code will directly accelerate drug discovery and improve patient outcomes worldwide

Cutting-Edge Work - Access to state-of-the-art AI infrastructure, enterprise LLM platforms, and rich scientific datasets

Continuous Learning - Dedicated learning budgets, access to AI research communities, and internal tech talks

Inclusive Culture - A diverse, global team united by the mission to Act for Patients

Competitive Compensation - Market-aligned salary, performance bonuses, and comprehensive benefits

Flexible Working - Hybrid work model with flexibility to balance deep work and collaboration

Career Growth - Clear Progression paths within Sanofi's engineering career framework, with opportunities to grow into L3-2 and beyond

Health & Wellbeing - Comprehensive health coverage, mental wellness programs, and employee assistance resources