Ayan Sau

About Ayan Sau

Senior Data Scientist / AI Specialist with 8+ years of experience delivering enterprise-scale AI solutions across Machine Learning, Generative AI, and data platforms. I build production-grade AI systems end-to-end—solutioning, architecture, microservices, deployment, and monitoring—focused on measurable business outcomes (growth, retention, efficiency, risk reduction).In my current role at Deloitte, I lead a 15-member team and own an Agentic AI platform and use cases from development to production. I built the platform initially using LangGraph and migrated it to Google ADK, implementing PostgreSQL checkpointing and LangSmith telemetry for deep observability—reducing debug time by ~30%. The system supports ~600 users globally (~60–67 requests/day) and delivers responses in ~15 seconds while integrating enterprise tools (Google Search, Azure DevOps) and safety controls (PII identification and moderation). Post-migration, UI development effort was reduced to zero by leveraging Gemini Enterprise’s in-house UI, with weekly sprint releases via Azure DevOps CI/CD.My work spans Agentic AI, RAG-based solutions, NLP/CV, and predictive modelling—solving complex problems like ERP cutover planning (40% effort reduction) and SDLC automation (30% faster cycles). I’m especially interested in responsible, scalable AI: LLMOps/MLOps, evaluation, governance, drift, explainability, and secure integration into enterprise ecosystems.I’m currently pursuing a doctorate focused on AI strategy, and I enjoy bridging advanced AI capabilities with pragmatic delivery—turning experimentation into reliable, adopted systems. Education & CertificationsPhD (Pursuing), AI Strategy – Swiss School of Business & Management, GenevaM.Tech, Data Science & Engineering – BITS PilaniMicroMasters, Statistics & Data Science – MITPG Cert, Data Science (Top Performer) – IIT GuwahatiMicrosoft Certified: AI Engineer | Data Scientist

Experience

1. Assistant Manager - Ai and Data Science

   Deloitte

   Nov 2025 — Present · Hyderabad, IN

   Architected and delivered an enterprise Agentic AI platform end-to-end (built initially on LangGraph, migrated to Google ADK), designing microservices, session management, and production-grade deployment patterns for multi-user workloads.• Implemented PostgreSQL checkpointing and LangSmith telemetry (agent/tool-level traces) to improve runtime transparency and operational reliability; reduced debug time by ~30% through trace-driven triage.• Delivered and migrated 3 production use cases (Marketing Workbench) — Branding, Campaign Strategy, Social Media Advertising — enabling adoption by ~600 users globally with ~60–67 requests/day.• Achieved ~15s response time while orchestrating across multiple enterprise data sources/tools; integrated Google Search, Azure DevOps, Deloitte PII Identifier, and Moderation API for secure enterprise usage.• Improved output quality with higher acceptance rate, improved relevance, and fewer escalations by instrumenting agent/tool responses and iterating using telemetry insights.• Led a 15-member team and owned the full operating model (development, deployments, monitoring, weekly sprint releases) using Azure DevOps CI/CD; reduced UI development effort to zero post-migration by leveraging Gemini Enterprise’s in-house UI.

Education

• Birla Institute of Technology and Science, Pilani

  Master of Technology - MTech, Data Science

  2020 — 2022

• Swiss School of Business and Management

  Doctorate in Data Science, Data Science, Artificial Intelligence and Machine Learning

• Massachusetts Institute of Technology

  MicroMaster's degree Field Of Study Statistics and Data Science , Statistics and Data Science (Completed Successfully)

• Indian Institute of Technology, Guwahati

  Post Graduate Certification in Data Science, Machine Learning and Deep Learning

• SRM IST Chennai

  Bachelor's degree, B.Tech ( Mechanical Engineering)