Job Description

Job Description
Seeking a professional to design and implement autonomous AI agents capable of handling multi-step scientific queries. The focus is on applications, integrating clinical data and scientific literature, ensuring safety and precision in the execution of complex research tasks.
Responsibilities:
Design and implement autonomous AI agents using frameworks such as LangGraph, CrewAI or AutoGen.
Develop reasoning loops (e.g., ReAct, Plan-and-Execute) that allow agents to decompose research objectives into executable subtasks.
Build and optimize RAG pipelines that integrate structured clinical data and scientific literature.
Create and maintain tools for AI agents, enabling secure interaction with Lifebit federated APIs, SQL databases, and bioinformatics execution engines.
Implement secure, isolated code interpreter capabilities to execute Python or R code for data visualization and statistical analysis.
Tune LLM models to improve accuracy in specific function calls and tool usage.
Develop evaluation frameworks that measure the performance, veracity and safety of agents in clinical contexts.
Work with security teams to ensure agents operate within strict data privacy limits, preventing leaks or unauthorized access.
Collaborate with product and UX teams to design intuitive interfaces that allow interacting with agent systems.
Scale agent workloads in production using Kubernetes.
Requirements:
Bachelor's or Master's degree in Computer Science, Artificial Intelligence, Machine Learning or quantitative field, PhD is valued.
2+ years in AI or ML engineering, developing validated products, preferably in biotech, healthcare technology, or product-focused SaaS.
Proficiency in Python, TypeScript and ML frameworks such as Langfuse, PyTorch, TensorFlow, JAX or Scikit-learn.
Work with language models, including tuning RAG architectures.
Familiarity with AWS, Azure or GCP and experience deploying models in Docker/Kubernetes.
Experience with biological, genomic or clinical data is valued.

Salary to receive
To agree