Job highlights

- Preferred Qualifications

Job Summary

The deepest hands-on engineering role in AI Engineering. You personally build the most complex AI solutions Cyber Architecture and Solutions delivers, lead the hardest engineering problems in the agentic framework, and develop the reusable components the rest of the team builds on.

Domain focus: AI Engineering (agentic framework build and AI solution engineering); hands-on engineering of complex AI solutions in production, building reusable components, libraries and patterns for agent and platform workstreams.

About the team

AI ML and Cloud Security is the function through which GSK is transforming security itself. The team designs and builds the agentic framework, a multi-agent platform moving security services across the Chief Security Office from rule-based to AI-driven operation. It owns AI architecture and AI engineering end to end: the agent layer, LLM orchestration, the AI and ML platform, and the data fabric pipeline that Cyber Architecture and Solutions designs, builds and tests before a managed service vendor operates it.

The team also governs AI tool security posture, advises on cloud security across the Azure and GCP estate, and is the front door between GSKs business units and RD and the technical AI function. AI security standards set here are binding on every AI initiative in GSK. Seventeen roles across the UK, India and the US.

What makes this role different

A build role for someone who writes production code every day and wants to keep doing so. Expect a technical interview covering LangChain or LlamaIndex, vector database selection and retrieval tuning, and how you would test an agent whose output is non-deterministic. Tutorial or proof-of-concept exposure will not be enough - we are looking for things you shipped and then maintained. There is no line management in this role, by design.

Responsibilities

This role will provide you the chance to lead key activities to progress your career. These responsibilities include some of the following:

- Personally build the most complex AI solutions Cyber Architecture and Solutions delivers, using LangChain, LlamaIndex,



vector databases and production Python.

- Build the shared SDK, evaluation harness and reusable components the agent and platform workstreams both depend on.

- Lead the hardest engineering problems in the agentic framework build.

- Develop reusable engineering components, libraries and patterns the wider team builds on.

- Build to the architecture specification and raise design problems early rather than engineering around them.

- Write tests and documentation to vendor handover standard.

- Contribute to CI/CD pipelines and engineering tooling for AI workloads.

- Integrate agents with Splunk, Microsoft Sentinel, CrowdStrike and ServiceNow - writing and maintaining the connectors, not specifying them.

- Mentor engineers and review their code properly.

- Bring one toolchain change a quarter to the team with a working spike and a migration cost, not a recommendation.

- Use approved AI tools in your day-to-day work to improve your own productivity and your teams ways of working, sharing what works and applying sound judgement, including validating outputs and following GSKs Responsible AI, privacy and security requirements.

- Design components and APIs within the solutions you build, and document the trade-offs behind each design decision so the rationale survives handover.

- Apply GxP and data integrity expectations where what you build touches validated systems.

Basic Qualifications

- 10+ years in AI or software engineering with a relevant technical degree - domain focus: AI Engineering (agentic framework build and AI solution engineering); hands-on engineering of complex AI solutions in production, building reusable components, libraries and patterns for agent and platform workstreams





- Agent build in production Python and Kubernetes - LLM integration, prompt-to-action engineering, LangChain or LlamaIndex orchestration, tool and function calling, and evaluation of non-deterministic output

- RAG pipeline engineering end to end - chunking and embedding strategy, vector database and index design, retrieval tuning and evaluation, hallucination and grounding checks

- LLM application security in code - direct and indirect prompt injection defences, output validation, secrets handling, tool-call allow-listing and audit logging

- Delivering to architecture specification with tests and documentation to vendor-handover standard - raising design problems back to architecture rather than engineering around them, and mentoring analysts

Preferred Qualifications

- CI/CD pipelines for AI and ML platforms

- LLM evaluation frameworks (Ragas, DeepEval or equivalent), fine-tuning and prompt optimisation

- AI red-teaming tools such as Garak or PyRIT

- Cloud platform engineering on Azure or GCP

- Code review of junior engineers work

- Typical tools and technologies: Production Python and Kubernetes, LangChain, LangGraph and LlamaIndex orchestration, vector databases and retrieval tuning across pgvector, Pinecone and Weaviate, model gateways such as LiteLLM, evaluation harnesses for non-deterministic output, reusable library and SDK design, CI/CD and automated testing.

Skills

- AI for Cybersecurity

- AI Safety Engineering

- Application Security

- Artificial Intelligence Techniques

- Cloud Security

- Computer Programming

- Data Security

- Identity Access Management (IAM)

- Information Systems

- Infrastructure Security

- LangChain (FrameWork)

- LlamaIndex

- LLM Evaluation

- ML Security

- Operating Systems Security

- Scripting

- Security Engineering

- Security Technologies

- Security Testing

- Vulnerability Management

Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.

📌 Principal - AI ML and Cloud Security (Bengaluru)
🏢 GSK India
📍 Bengaluru