About the Role
We are looking for an AI Developer to play a key role in building our agentic reasoning engines. You will own the intelligent layer of our bespoke enterprise system, ensuring our AI agents are reliable, secure, and capable of solving complex workflows. You will be involved in developing sophisticated RAG capabilities and stateful multi-agentic networks, moving beyond simple chat to high-consequence autonomous systems.
What You’ll Do
- Design and implement stateful multi-agent networks and/or workflows using LangGraph and/or LangChain.
- Build and optimise end-to-end RAG pipelines, focusing on high-precision retrieval, semantic search, and the integration of diverse data sources (vector DB, graph DB, RDBMS, etc).
- Architect and implement multi-layered guardrails to ensure agent actions remain within business scope, enterprise safety and policy boundaries.
- Build and maintain high-performance AI microservices, ensuring they are optimised for OCI-compliant environments.
- Use of LLM evaluation frameworks to quantitatively measure agent performance.
- Partner with software teams to define data contracts and integrate information flow from AI layer to software backend and frontend.
Qualifications
Technical Requirements
- Expert-level proficiency in Python, specifically for asynchronous AI applications.
- Mastery of LangGraph and LangChain for building production-grade, stateful systems with human-in-the-loop verification patterns.
- Proven success in developing RAG capabilities, including experience with chunking strategies, advanced techniques like query expansion, re-ranking, hybrid search, etc.
- Deep hands-on experience with vector DB, graph DB, NoSQL and RDBMS.
- Experience building guardrails to mitigate risks such as prompt injection and data leakage.
- Experience integrating AI evaluation into CI/CD pipelines (eg, Jenkins, GitLab CI).
- Experience serving AI agents via FastAPI, packaging and deploying in OCI-compliant environments and OCP.
- Proficiency in AI tracing, observability and logging tools to identify and fix bottlenecks in complex reasoning paths.
- Experience implementing semantic caching and designing for parallel LLM invocation.
- Knowledge of prompt engineering and context engineering techniques.
Bonus
- Working knowledge of database query languages including SQL and HiveQL (or equivalent).
- Working knowledge of building data ETL pipelines.
- Knowledge in building or fine-tuning deep learning models.
- Knowledge in mathematical statistical analysis.