Designing and implementing agentic AI software systems.

Working with the Agentic Lab on researching RAG mechanisms to improve LLM forecasting capabilities.

1. Teaching assistant for Computer Vision course (Prof. Davi Geiger). Designed detailed assignments on CNNs, Vision Transformers, and Object Tracking for 40 undergraduate student.

2. Evaluated student implementations of state-of-the-art computer vision architectures and provided technical mentorship on deep learning best practices

1. Migrated outsourced marketing reporting to a Snowflake analytics layer powering 14 Power BI dashboards; worked as part of cross-functional team to eliminate outsourcing vendor spends.

2. Developed 8 production-grade Snowflake SQL stored procedures (avg. 200+ LOC) and 4 core analytics layer

dimension/fact tables; delivering 4-6 Power BI dashboards aligned to marketing stakeholder KPI definitions

3. Automated SQL deployments with Azure DevOps CI/CD: ~3 hours to <20 minutes, saving 60+ engineering-hours per sprint cycle.

1. Reduced merchant onboarding drop-off by 40% by building real-time sales funnel dashboards and identifying critical leakage points in KYC verification workflow, enabling product team to prioritize UX improvements that increased conversion

2. Developed production OCR pipeline achieving 97% accuracy in extracting/validating business document numbers using Google Vision APIs and OpenCV. Deployed via REST endpoint (AWS Lambda) and integrated into merchant onboarding workflow, eliminating manual entry requirements

3. Created RAG-based conversational agent using LangChain and Milvus vector database to automate customer query resolution through agent-tool workflows, handling 1K+ daily queries with 85%+ resolution rate