Ankit Sharma

AI/ML Projects & Leadership Experience

1. InsightX – Generative AI/ML Experience Platform

• Conceived and led end-to-end development of a scalable GenAI platform.
• Defined AI strategy to accelerate pre-sales and demonstrate AI/ML capabilities.
• Reduced solutioning time and elevated firm's AI leadership in client engagements.

2. Personal Information Retrieval System (PML Compliance)

• Designed a multi-agent AI system to extract summaries and sensitive firm/client data.
• Automated PML review recommendations, reducing manual review time by 60%.
• Led secure production deployment using the Azure ecosystem and Prompt Flow.
• Tech Stack: Python, FastAPI, Azure Storage, Cosmos DB, GitHub Actions, APIM, Entra ID.

3. AI-Powered Email Automation for Logistics

• Built a multi-agent system to classify emails, extract tracking info, and generate responses.
• Streamlined shipment communication and reduced manual processing efforts.
• Tech Stack: Google GenAI, LangChain, LangGraph, Composio, Playwright, GCP, Docker.

4. Sawmill Workflow Optimization – Predictive Maintenance

• Led predictive modeling initiative to optimize machine maintenance schedules.
• Mentored team members and ensured timely delivery of business-aligned milestones.
• Tools: Microsoft Fabric, Machine Learning.

5. Vehicle Manuals Q&A System – RAG Implementation

• Applied advanced Retrieval-Augmented Generation (RAG) techniques.
• Enhanced contextual search and Q&A accuracy in automotive manuals.

1. Fraud Detection & Entity Resolution Models – TransUnion

• Developed and deployed load fraud detection, entity resolution, and DOB confidence scoring models.
• Delivered $1M+ business impact through enhanced risk detection and operational efficiency.

2. Multi-Agent Text-to-SQL Framework – TransUnion Ledger

• Designed a multi-agent Generative AI-powered Text-to-SQL framework for querying TransUnion Ledger databases.
• Enabled faster and more accurate access to complex financial datasets across business units.

1. Personalized Recommendation Engine – Fashion Retail

• Engineered a personalized recommendation engine for a leading fashion retail client.
• Enhanced customer engagement and improved conversion rates through tailored product suggestions.

2. Anomaly Detection System – Oil & Gas

• Implemented an anomaly detection algorithm using autoencoders to identify early signs of equipment failure.
• Reduced downtime and maintenance costs by enabling proactive issue detection.

• Automated financial workflows and reporting using Python, Sikuli, and NLP techniques.