About the Role :

You'll lead Vecton's AI pod embedded inside a financial institution's engineering organisation - a team of forward-deployed AI Engineers building agentic systems for the client's software development lifecycle, plus AI applications for their business use cases.

You are the pod's architect and its most senior engineer. You decide how solutions get built, you build the hard parts yourself, and you're accountable for everything the pod ships.

This is a hands-on role : Roughly 60 - 70% of your time is engineering - architecture, the critical path, the parts nobody else on the pod can do yet. The remaining is planning, prioritisation, review, and client coordination. If you're looking for a role where you stop writing code, this isn't it.

You'll work closely with Client Engineering and Product Leaders on project planning and sprint definition, with a dotted line to the client's technical leadership.

What You'll Own :

Architecture & Technical Direction :

- Own the solution architecture for everything the pod builds - agentic SDLC systems and AI applications alike.

- Choose the stack : agent frameworks vs. established SDKs (Claude Code, Agent SDKs), orchestration patterns, retrieval strategy, model selection, eval approach - and defend those calls to Vecton and to the client.

- Design the context layer : how agents get codebase insight, integration docs, client standards, and prior decisions at the right moment.

- Set the engineering standards the pod works to - repo structure, eval harnesses, prompt/tool versioning, deployment patterns.

Hands-On Build :

- Personally build the critical path : the hardest agent, the orchestration spine, the piece with the most technical risk.

- Prototype fast when a direction is unproven - de-risk before the pod commits sprint capacity to it.

- Stay deep enough in the code that your reviews are real reviews.

Team & Delivery :

- Review the pod's work - architecture, code, prompts, agent behaviour, evals.

- Guide FDEs technically : unblock them, correct direction early, raise their ceiling.

- Be accountable for what the pod delivers - quality, timelines, and whether it actually gets used.

- Run sprint planning and prioritisation.

Leadership Alignment :

- Work with client engineering leaders to identify and qualify use cases.

- Translate client problems into scoped, sequenced technical work.

- Present architecture and solution design to client architects, security, and technical leadership.

What We're Looking For :

- 6-10 years of software engineering experience, including senior/lead ownership of production systems end to end.

- Deep fundamentals - system design, API design, testing, CI/CD, debugging production issues at scale. You need these to build agents that operate on real codebases.

- Strong hands-on AI engineering : LLMs, prompt engineering, RAG, agentic frameworks (LangChain, LangGraph, CrewAI, AutoGen), vector databases and embedding pipelines.

- Track record of shipping AI systems to production - not experiments.

- Python depth and real working knowledge of at least one full stack (MERN/MEAN or Java/Spring Boot).

- Experience leading a small engineering team.

- A clear, defensible point of view on where AI genuinely accelerates the SDLC and where it doesn't - you'll be asked to justify it by client architects who are sceptical by default.