• Production RAG: indexing, retrieval, hybrid search, reranking, query rewriting, grounding, citations • Context Graph: entity resolution + linking + provenance; graph + vector retrieval; supports multi-hop context • LLM orchestration: tool/function calling, structured outputs, routing across model tiers, failure modes • GPU/inference cost optimization: batching, caching/KV reuse, quantization, autoscaling; optimize $/session + latency • Safety + compliance: PII/PHI handling, redaction, audit logs, deterministic replay, hallucination mitigation • LLMOps: eval harness (golden sets, regression, adversarial), monitoring for quality/cost/drift • Design/ship the end-to-end pipeline: retrieve → assemble context → generate → cite → log/monitor • Improve quality and trust via evaluation, feedback loops, and clear evidence-backed outputs • Partner with product, security, and domain teams; write crisp design docs; raise engineering bar • Ship RAG v1 with citations + measurable quality metrics • Deliver Context Graph v1 that improves retrieval on real copilot tasks • Reduce cost/latency with a concrete inference optimization plan shipped to prod.

• Build and deploy enterprise-ready AI solutions using LLMs, other GenAI approaches, and deep learning capabilities. • Scale out AI solutions: optimize performance of solutions, automate deployment and testing • Rapidly prototype and iterate on AI applications using Azure, AWS, and off-the-shelf tools. • Partner with the AI Solutions Architect to ensure scalable, secure, and compliant system design. • Develop APIs and lightweight UIs (e.g., Dash, Flask and others) to deliver AI tools to end users. • Stay current on emerging AI technologies, including vector databases, RAG pipelines, and productivity AI platforms. • Drive delivery of AI components aligned with product roadmaps and business priorities.

• Design, build, and operate LLM-powered systems using Gemini and Vertex AI • Implement RAG architectures at scale, including ingestion, retrieval, and generation • Build and orchestrate LLM agents using LangChain or similar frameworks • Integrate AI capabilities via API-driven architectures • Debug and optimize end-to-end LLM pipelines: Chunking strategies, Embeddings, Retrieval logic, LLM response behavior • Deliver production-ready AI services, including: Monitoring and observability, Rate limiting and cost controls, Reliability and fallback strategies • Contribute to solution design and technical decision-making • Continuously evaluate and experiment with new LLM models and platform features • Implement AI safety, security, and compliance controls • Collaborate with cross-functional teams across time zones

• Agentic Engineering & Orchestration Workflow Design: Architect complex, multi-agent workflows using Microsoft AI tech stack. • Design, Develop and Deploy agents to handle loops, interruptions, and human-in-the-loop interventions. • Tool Use & Function Calling: Build reliable "tool layers" that allow LLMs to safely interact with internal APIs, databases, and third-party SaaS platforms (e.g., Salesforce, Workday, ServiceNow etc.). • State Management: Design persistence layers to manage agent memory, conversational history, and context windows efficiently. • Advanced Data & RAG Strategy Retrieval Pipelines: Build production-grade data retrieval and integration systems. • Optimize vector indexing, document chunking, and re-ranking algorithms to ensure high-precision context retrieval. • Data Quality: Collaborate with Data Engineers to curate "Golden Datasets" for agent consumption. • LLMOps, Evaluation & Quality Automated Evaluation: Build CI/CD pipelines for AI that include "LLM-as-a-Judge" testing. • Leverage frameworks to score agent outputs for accuracy, hallucination, and safety before deployment. • Observability: Instrument applications with tracing tools to visualize agent reasoning chains, monitor latency, and debug failures in production. • Cost Optimization: Monitor token usage and latency, optimizing prompt density and caching strategies to maintain high performance at sustainable costs. • Innovation & Collaboration: Prototyping to Production: Rapidly validate new ideas using state-of-the-art models, then refactor successful prototypes into maintainable, tested production code. • Standards Adoption: Stay ahead of the curve by evaluating emerging technologies to standardize agent connectivity.