Technical PM- AI

Technical Delivery Lead / Tech PM - AI Solutions POD (ML + Gen AI, RAG)

Exp: 8 - 10+ years in technology delivery, including AI/ML initiatives

We need a hands-on Technical Delivery Lead/Tech PM to drive delivery for an AI solutions Pod building production-grade AI capabilities (ML and Gen AI), including RAG-based solutions.

The role requires someone who can plan and deliver based on component interactions (data --> embeddings/vector stores --> retrieval --> LLM orchestration --> evaluation --> deployment), set guardrails, and guide decisions on when ML/Gen AI is appropriate vs when it isnt.

This person should be trusted to represent the pod in senior forums and be able to run delivery across multiple pods/squads when needed.

Key responsibilites:

• Own end-to-end delivery of AI based solutions: roadmap, milestones, dependency management, delivery governance across 1-2+ pods
• solution planning based on architecture: create delivery plans that reflect how components integrate (data ingestion, vectorization, retrieval, model endpoints, orchestration, UI/API,monitoring)
• Stakeholder Leadership: represent the team in architecture reviews, governance and senior stakeholder updates, provide crisp reporting and decision options.
• Delivery excellence: manage risks, NFRs (latency, resilience, security), release planning, production readiness, incident learnings
• Technical oversight of ML+Gen AI:

Differentiate and select approaches: classical ML vs Gen AI vs hybrid patterns (eg: RAG + ML ranking / classification)

Define where ML adds value (prediction, scoring, classification) and where Gen AI adds value (generation, summization, extraction, conversational interfaces)

• RAG delivery Leadership

Chunking strategies, embedding model selection, indexing, retrieval patterns, reranking, citation/attribution, freshness updates

work with teams on relevance evaluation and hallucination reduction patterns

• Guardrails and controls

Define Guardrails for data usage, sensitive data handling, access controls, content safety, prompt/response filtering, and human-in-the-loop where required

Drive policies/standards for model usage, tool access, logging, monitoring, and approval gates

Must have experience and capabilities:

• 8-10+ years in technical delivery / engineering-led programme delivery
• Proven delivery of multiple AI use cases into production (not only POCs)
• Hands-on technical comfort: able to work with engineers on design decisions, challenge approaches, and translate requirements into implementable epics/stories
• strong understanding of

ML Lifecycle (data prep, training/validation, bias considerations, evaluation, deployment, monitoring drift)

GenAI Lifecycle (model selection, orchestration, prompt strategies at a high level, evaluation, safety)

RAG Patterns (vector stores, embeddings, retrieval, reranking, grounding, citations)

• Working knowledge of cloud architecture and CI/CD for AI systems (MLOps/ LLMOps concepts)
• Solid grounding in data engineering concepts: data quality, data lineage, metadata, batch/stream ingestion, access controls.

Preferred skills:

• Experience with model risk / governance in regulated environments
• Familiarity with vector databases and search (eg: Elasticsearch/opensearch vector, Pinecone, Weaviate, pgvector - depending on bank standards)
• Ability to define and track AI KPIs: accuracy/relevance, latency, cost per request, adoption, defect leakage, incidents

What success looks like (outcome-based)

AI solutions shipped with measurable business impact, not just demos

clear, realistic plans that account for data readiness + integration complexity

Guardrails and controls in place so solutions are safe, compliant, and supportable

Stakeholders trust the lead to make decisions and keep delivery moving

If you are keen to join us, you will be part of an organization that values your contributions, recognizes your potential, and provides ample opportunities for growth.

For more information, visit www.capco.com.

Follow us on Twitter, Facebook, LinkedIn, and YouTube.