About the RoleWe’re seeking an AI Software Development Lead to spearhead AI-assisted software development adoption across BFSI projects and lead solutioning for client proposals and pre-sales engagements.Will champion vibe coding—the emerging practice of using LLMs and coding agents (e.g., GitHub Copilot, Cursor, Claude Code, etc.) to generate working code from natural-language instructions, iterating rapidly while enforcing quality and compliance. Your leadership will modernize engineering workflows and scale AI-first development practices across diverse BFSI portfolios.Will architect and deliver enterprise-grade AI applications leveraging Generative AI (GenAI), Agentic AI, LLMs, RAG, and Agentic RAG—with a strong focus on security, governance, observability, and cost efficiency.This role operationalizes AI-first delivery, increases developer productivity, strengthens proposal win rates through compelling AI solutioning, and ensures secure, compliant implementations aligned with BFSI standards.Key Responsibilities
- AI-Assisted Development Leadership
- Drive organization-wide adoption of coding agents and vibe coding practices; define guardrails, standards, and governance for BFSI environments.
- Build playbooks for prompt engineering, code generation, refactoring, test generation, documentation, and secure patterns using Copilot/Cursor/Claude Code, etc.
- Deliver enablement programs: workshops, hands-on labs, brown-bags; establish usage analytics and productivity KPIs.
- Solutioning, Pre-Sales & Proposal Support
- Partner with sales, pre-sales, service lines, and delivery to:
- tailor AI-first roadmaps, demo assets, and POCs/Pilots
- lead technical solutioning for RFPs/RFIs: author architecture options, reference designs, delivery models, and cost estimates.
- Create client-facing proposals with clear business outcomes, risk/compliance alignment, and measurable success metrics.
- Architecture & Delivery (LLMs, RAG, Agents)
- Architect and deliver agentic systems—tool orchestration, planning/critique loops, memory, multi-agent collaboration for complex BFSI workflows.
- Own end-to-end solutioning: data acquisition/transform; embeddings/retrieval; prompt pipelines; function calling/tool schemas; APIs/SDKs; UI integration.
- RAG & Agentic RAG Best Practices
- Design advanced RAG pipelines: chunking, hybrid retrieval (vector + keyword), rerankers, query rewriting, context compression, caching, grounding, and citations.
- Build Agentic RAG flows combining retrieval + tool use + planning loops to maximize accuracy, policy adherence, and cost performance.
- Quality, Evals & Observability
- Define LLM/agent evaluation: groundedness, factuality, precision/recall, hallucination rate, agent success rate, latency, cost/query.
- Implement observability: tracing, token/cost accounting, prompt/version lineage, user feedback loops, and red-team logs.
- Collaboration & Leadership
- Mentor engineers; lead design reviews and AI SDLC standards; influence architecture councils.
- Drive build-vs-buy decisions, vendor evaluations, and cost/latency optimization strategies.
Required Qualifications
- 10–12+ years in software engineering, including enterprise architecture and delivery; 3+ years hands-on with LLMs/GenAI.
- Proven BFSI exposure (banking, payments, insurance, capital markets, fintech) with security/compliance constraints.
- Strong software engineering fundamentals and coding in Python plus one of Java/TypeScript/C#/Scala; production APIs/microservices; CI/CD.
- Hands-on with coding agents & IDEs: GitHub Copilot, Cursor, Claude Code etc. and IDE integrations (VS Code/IntelliJ/JetBrains); expert in vibe coding workflows.
- Deep knowledge of LLM ecosystems (e.g., Azure OpenAI/OpenAI, Anthropic, Google, Meta): prompting, function calling, MCP (Model Context Protocol), token/cost management.
- Expertise in RAG: vector DBs (FAISS, Pinecone, Milvus, pgvector/Postgres, Elastic/OpenSearch), embedding strategies, and rerankers.
- Experience with agent frameworks: LangGraph, AutoGen, Sematic Kernel, CrewAI (or similar) and tool integrations.
- LLMOps and eval tooling: LangSmith, TruLens, Ragas, DeepEval, W&B, MLflow; prompt caching/compression; distillation.
- Infra & data: Docker/Kubernetes, Azure/AWS/GCP, Kafka, Airflow; API security and secrets management.
- Testing & quality: unit/integration/e2e, canary/blue-green, A/B or interleaving experiments for AI features.
- Excellent communication; able to translate BFSI needs into reliable AI systems with clear KPIs.
Education & Certifications
- Bachelor’s/Master’s in Computer Science, Software Engineering, Data/AI, or related field (or equivalent experience).
- Preferred: Cloud architect/ML/AI certifications (Azure/AWS/GCP)
