Aug 2024 – Dec 2025 · 1 yr 4 mos

• 🚀 Driving productivity & growth through LLMs, Agentic AI, and Automation
• Validation Automation Framework – Designed & implemented a Playwright MCP-based framework with self-healing test generation (GitHub Copilot), integrated with Azure DevOps & Jira, cutting QA effort by 60%.
• Legal AI & Copilot Platform – Led 15-member team delivering AI-powered legal automation (document comparison, Gen AI bots, M365 workflows) processing 200+ GB data for 500+ users, saving 2–3 hours weekly per lawyer.
• Responsible AI Compliance Agent – Ideated, architected & designed an enterprise multi-agent governance assistant for LLM safety, pricing guidance & risk controls (prompt injection, jailbreak, cross-prompt). Automated RAI test cases & policy feedback, reducing compliance review effort by 20%+.
• AI Data Lens – Built & scaled with a team of 5+, a natural language assistant for structured data (Power BI, Fabric Lakehouse) enabling SQL/DAX generation, anomaly detection & narrative insights—accelerating business decision cycles.
• DevelopFAST Accelerator – Conceived & authored core prompts & high-impact use cases for a multi-agent Gen AI SDLC accelerator, automating requirements, code, tests & architecture reviews, delivering ~30% faster delivery.
• Video Generation Agent – Designed architecture & solution automating content creation (Playwright recording, scene detection, narration, subtitles, B-roll) reducing video production effort by 25%.
• Communication Agent – Built Outlook add-in for smart formatting, templates & spell-check, improving email productivity by 20%.

Mar 2023 – Aug 2024 · 1 yr 5 mos

• 📊 Managing Data Engineering & Data Science portfolios | Building enterprise OpenAI-powered applications
• RAG-based Chatbots – Led projects delivering large-scale OpenAI RAG chatbots, enabling knowledge retrieval and conversational automation across enterprise datasets.
• Event AI Assistant – Managed a 5-member team to engineer an Azure OpenAI–powered assistant for 250K global users, handling 2,000+ requests/min with load balancing, intent classification, 10+ data sources, and hybrid session recommendations.
• Training Agent – Built a pipeline converting MP4/PPT/Text/Docs into quiz-based modules, expanding training content coverage by 80%.
• Automated Newsletter – Designed a multi-source summarization system that reduced manual content curation effort by 80%.
• Interview Automation – Developed a Gen AI portal for resume screening, dynamic question generation & candidate filtering across tech stacks, streamlining recruitment workflows.

Mar 2022 – Feb 2023 · 11 mos

• 📈 Forecasting & Predictive Modeling
• Retail Demand Forecasting – Led a 3-member team delivering 1M+ weekly bottom-up Random Forest forecasts with automated validation, retraining & monitoring; improved inventory & labor planning by 20% and achieved MLOps maturity level 4.
• Cash Flow Forecasting – Cut monthly forecast error from 20% → <5% by leveraging synthetic data generation, stacked time-series models, meta-learners & imbalanced classification, strengthening liquidity & capital planning.

Apr 2021 – Feb 2022 · 10 mos

• ⚖️ Legal & Data Engineering Solutions
• Legal Risk Classifier – Developed an ensemble ML model that improved F1 score by 50%+ on imbalanced data, automating the scanning of 2M+ tickets/day and cutting manual review workload by 50%.
• ETL Migration – Led a 10+ member team in migrating Qlik pipelines to SQL, processing 100+ GB/day, significantly improving data stability, accessibility, and scalability.

Oct 2019 – Mar 2021 · 1 yr 5 mos

• 🛠️ AI for Support & Text Analytics
• Escalation Risk Prediction Model – Built a ML model scoring 1M+ tickets/hour across 10M+ sparse support records, delivering 40%+ productivity gains and accelerating global issue resolution.
• Unsupervised Topic Generation Engine – Applied clustering on retail, events, code comments & social feedback datasets, reducing manual analysis by 20%+ and enabling faster insights.
• Ticket Similarity Recommendation Service – Developed an embeddings-based recommendation engine with Flask APIs, reducing ticket resolution time by 25% and improving support team efficiency.
• Text Analytics Engine – Core contributor to backend ML APIs for topic modeling, keyphrase extraction, sentiment & PII detection; deployed as SDK + Azure Marketplace offerings.

Aug 2017 – Sep 2019 · 2 yrs 1 mo

• 💡 Early AI & Data Engineering Projects
• CRM ETL Pipeline – Engineered a scalable ETL for opportunity & lead data, enabling faster downstream analytics and improved sales insights.
• Media Analytics Solution – Contributed to research & implementation of face recognition and emotion detection models; delivered as a packaged Azure Marketplace solution.
• Ticket Similarity Recommendation Service – Designed an embeddings-based recommendation engine (Flask APIs) that reduced ticket resolution time by 25% and boosted support efficiency.

May 2016 – Jun 2016 · 1 mo · Kharagpur Area, India

• I worked here as a research intern under Complex Networks Research Group, IIT Kharagpur.
• Developed a locally weighted regression model for surcharge & fare prediction (MSE 0.01), enabling dynamic trip cost comparison.