This is us
At Avenga, we believe that human creativity empowers technology that matters. Operating globally, our 6000+ specialists provide a full spectrum of services, including business and tech advisory, enterprise solutions, CX, UX, and UI design, managed services, product development, and software development.
This is the job
We are seeking a Senior AI Engineer to drive the development of AI business agents using the SADK framework. Ideal candidates bring deep expertise in AI/ML/LLMs, RAG pipelines, vector databases, and agent frameworks, with a preference for healthcare/clinical trials experience and compliance knowledge.
The project is centered on the implementation of clinical solutions for pharmaceutical product development and testing.
This is you
- 5+ years in software engineering, including 3+ years with AI/ML/LLMs;
- Expert-level prompt engineering and LLM application development;
- Experience with OpenAI, Anthropic, or Azure OpenAI APIs;
- Strong proficiency in Python and/or TypeScript;
- Hands-on RAG (Retrieval Augmented Generation) implementation;
- Understanding of agent frameworks like LangChain or AutoGen;
- English - Upper-Intermediate level or higher.
Nice-to-have skills:
- Domain knowledge in healthcare or clinical trials;
- Experience with Microsoft Agent Framework or similar tools;
- Skills in fine-tuning and model optimization;
- Background in multi-agent systems and workflows;
- Knowledge of 21 CFR Part 11 and GxP compliance.
This is your role
- Lead development of business agents using the SADK framework;
- Design agent architectures, including Knowledge, Review, Triage, Workflow, Research, and Recommendation agents;
- Master prompt engineering techniques tailored to clinical use cases;
- Implement RAG pipelines and integrate knowledge bases;
- Optimize agent performance for accuracy, latency, and cost;
- Establish evaluation frameworks and quality metrics for agents;
- Mentor AI/ML engineers on best practices in agent development;
- Collaborate with domain analysts to translate requirements into agent specifications;
- Partner with the Platform Pod to improve SADK and identify new MCP server needs.