Job Title: Lead Technical Consultant – Generative AI (GenAI)
Experience
- 6–8 years in software engineering / AI engineering / solution engineering
- 3+ years hands-on experience delivering GenAI solutions on cloud (AWS/Azure/GCP) to production
Role Summary
As a Lead Technical Consultant – GenAI, you will own end-to-end delivery of production-grade GenAI solutions—covering architecture, build, deployment, evaluation, observability, security, and cost controls. You will work closely with business stakeholders, SMEs, and engineering teams to convert ideas into scalable solutions using RAG, agents, structured extraction, and workflow automation.
Key Responsibilities
Solution Design & Architecture
- Design GenAI system architectures for enterprise use cases: RAG, agentic workflows, document intelligence (IDP), copilots, conversational assistants, and decision support.
- Define architecture for LLM orchestration, tool/function calling, memory, prompt management, and retrieval strategies.
- Translate non-functional requirements into design: latency, throughput, concurrency, availability, compliance, and data isolation.
Hands-on Development (Core Expectation)
- Build and ship solutions using Python (must-have), and relevant GenAI frameworks (e.g., LangChain/LangGraph/CrewAI/Semantic Kernel or equivalent).
- Implement retrieval pipelines: chunking, embeddings, vector DB setup, hybrid search, reranking, metadata filtering, caching.
- Build structured output pipelines: JSON schema extraction, validation, reconciliation, deterministic post-processing, and HITL loops.
Production Readiness & LLMOps
- Implement and manage evaluation frameworks: offline/online evals, golden datasets, regression testing, LLM-as-judge + human review.
- Set up monitoring & observability: quality metrics, drift detection, prompt/version tracking, latency, cost/token usage dashboards.
- Drive release management practices: CI/CD for prompts + code, environment promotion, rollback strategies.
Security, Governance & Responsible AI
- Apply security best practices: PII handling, secrets management, RBAC, network isolation, encryption, and audit logging.
- Implement guardrails: prompt injection resistance, data exfiltration prevention, safety policies, output constraints, citation/grounding checks.
- Ensure compliance with enterprise policies and client requirements.
Stakeholder Management & Leadership
- Own technical delivery for 1–2 workstreams; break down work into milestones/sprints.
- Mentor junior engineers; lead code reviews, design reviews, and engineering best practices.
- Partner with SMEs for validation workflows and continuous improvement of accuracy.
Must-Have Skills
- Strong software engineering foundation with Python (fastAPI/flask, async patterns, testing, packaging).
- 3+ years implementing GenAI solutions using commercial/open LLMs (Azure OpenAI / OpenAI / Bedrock / Vertex / OSS).
- Proven experience with RAG patterns and vector databases (Pinecone/FAISS/Weaviate/Chroma/Elastic/OpenSearch).
- Solid cloud experience (AWS/Azure/GCP): compute, storage, networking, IAM, monitoring.
- Experience building APIs/microservices and integrating with enterprise systems.
- Practical experience with evaluations, prompt/versioning, and production troubleshooting.
Good-to-Have Skills
- Document AI / IDP: OCR, layout understanding, table extraction, reconciliation logic.
- Agentic systems: tool routing, planner/executor patterns, multi-agent orchestration.
- Data engineering basics: pipelines, ETL, data quality, metadata-driven ingestion.
- Containers & orchestration: Docker, Kubernetes.
- Familiarity with governance frameworks and secure enterprise deployments.
Qualifications
- Bachelor’s degree in Computer Science / IT / Engineering (or equivalent practical experience).
- Certifications (nice to have): AWS/Azure, GenAI specialization, Kubernetes, security.
Design GenAI system architectures for enterprise use cases: RAG, agentic workflows, document intelligence (IDP), copilots, conversational assistants, and decision support.· Define architecture for LLM orchestration, tool/function calling, memory, prompt management, and retrieval strategies.· Translate non-functional requirements into design: latency, throughput, concurrency, availability, compliance, and data isolation.Hands-on Development (Core Expectation)· Build and ship solutions using Python (must-have), and relevant GenAI frameworks (e.g., LangChain/LangGraph/CrewAI/Semantic Kernel or equivalent).· Implement retrieval pipelines: chunking, embeddings, vector DB setup, hybrid search, reranking, metadata filtering, caching.· Build structured output pipelines: JSON schema extraction, validation, reconciliation, deterministic post-processing, and HITL loops
Degree