Modality: Hybrid (3 days in office, 2 days remotely)

Location: Cambridge, MA or Gaithersburg, MD

What You’ll Do

• Design, build, and operate the agentic and LLM-powered systems that biologics scientists rely on, owning them from concept through reliable, monitored production use. 

• Build custom agentic skills and tools that encode our scientists’ expertise, so an agent can do real domain work rather than generic chat. 

• Connect agents to internal data, models, and services through AZ’s approved integration platforms, in line with AZ data and security standards. 

• Build and deploy LLM and agentic applications end to end, including the interfaces and the supporting engineering (authentication, logging, evaluation, error handling) that makes a tool dependable. 

• Build the digital pipelines that move data between our models, our high-throughput and automated lab platforms, and the scientists who use the results. 

• Partner day to day with protein scientists, computational biologists, and platform engineers, and work in close partnership with BIX, EAI, and R&D IT, reusing shared platforms and data products rather than building in isolation. 

• Document your work in GitHub and Confluence so others can maintain and extend it, and meet relevant safety, quality, and compliance standards, including FAIR (Findable, Accessible, Interoperable, Reusable) data practices. 

How You’ll Work 

You will work in the CLI or the coding environment you prefer, such as VS Code, using code assistants and agents day to day to write and ship code more efficiently. Our stack includes Claude Code, Claude Cowork, GitHub Copilot, M365 Copilot, and HuggingFace, with GitHub and Confluence for code and documentation. You will deploy on AZ infrastructure including scientific computing platforms, AWS, Kubernetes, and Domino, with access to high-end GPU compute (including AZ’s sovereign AI compute platform built on NVIDIA DGX SuperPOD). 

What You Bring 

Essential Education & Experience 

• MS degree in Computer Science, Software Engineering, Computational Biology, Data Science, or a related quantitative field, or equivalent demonstrated ability. 

• 3–10 years of relevant software engineering experience. 
• Proven, hands-on experience building agentic AI systems and deploying them to production. We look for something you personally built and shipped that people used, whether at a company, a startup, a lab, or a substantial open-source or personal project. 

Essential Skills 

• Solid Python and/or TypeScript programming skills, and hands-on experience in software development best practices: clean code, version control, testing, and the judgment to build something that keeps working after you have moved on. 

• Practical experience building with LLMs and agent frameworks, including tool and function calling, retrieval, and orchestration of multi-step workflows. 

• Fluency with modern agentic developer tools such as Claude Code, Copilot, or similar, used daily as part of how you build. 

• Clear written and verbal communication, with the ability to explain technical choices to scientists who are not engineers. 

• Working knowledge of Unix, SQL databases, REST APIs, and cloud computing (AWS). 

Desired Skills 

• Experience with tool-integration layers that connect agents to enterprise data, models, and services. 

• Experience with AI frameworks such as Pydantic-AI, LangChain, or similar. 

• Experience building production retrieval-augmented generation (RAG) pipelines and working with vector databases. 

• Experience in at least one compiled programming language (e.g., C, Java, Go, or Rust). 

• Experience with the evaluation and guardrails that make LLM applications trustworthy, and cloud deployment practice: containers, Kubernetes, CI/CD, and platforms such as Domino. 

• Familiarity with laboratory automation, high-throughput platforms, or scientific data, and awareness of responsible-AI and compliance considerations in a regulated environment. A life-sciences background is not required, but curiosity about biology and drug discovery is. 

Who You Are 

You learn quickly, care about engineering craft, and communicate clearly with scientists. You will work close to the science and close to your users, alongside teams spanning BE, BIX, EAI, and R&D IT, on work that amplifies our science and helps bring the right medicines to patients faster.

What You’ll Accomplish in Your First Year

First 30 days. Get hands-on with our stack, including our internal AI platforms, agent frameworks, and the custom tools already in use, and ship your first improvement. Sit with the scientists you build for to learn where they lose time. Scope a first use case end to end with your manager.

First 90 days. Deliver production agentic tools that scientific teams use in their day-to-day workflows. Working with BIX and R&D IT, connect an approved internal data source, such as assay data, to our agent tooling through AstraZeneca's (AZ) governed integration platforms. Establish the practices that make agent code dependable: evaluation, guardrails, and human oversight at the right decision points.

First 180 days. Own a small portfolio of shipped tools with measurable adoption and time saved. Help build the digital pipelines that connect our data and models to high-throughput lab platforms. Contribute reusable components to BE's shared skills repository so the next builders inherit your work.

The annual base pay for this position ranges from $137,349.40 to $206,091.60 annually. Our positions offer eligibility for various incentives—an opportunity to receive short-term incentive bonuses, equity-based awards for salaried roles and commissions for sales roles. Benefits offered include qualified retirement programs, paid time off (i.e., vacation, holiday, and leaves), as well as health, dental, and vision coverage in accordance with the terms of the applicable plans.

Date Posted

Closing Date