At Salesforce, I build enterprise AI agents, high-throughput data reconciliation engines, and full-stack platforms that automate complex developer workflows.
I built an autonomous Salesforce provisioning agent using LangChain and Python, combining ReAct prompting, Salesforce APIs, synthetic data generation, and RAG validation. It reduced environment setup time by 85%, achieved a 98% bootstrapping success rate, and eliminated metadata deployment failures caused by governor limits.
I also architected a reconciliation framework using Levenshtein matching and Annoy vector search, reducing redundant insertions by more than 95% and duplicate lookup latency to under 10ms. For Salesforce solution building, I created a React, Flask, and Python platform that reduced sandbox-to-repository extraction from two hours to under two minutes.
Previously, I interned at Salesforce building a graph engine for navigating custom-entity relationships, and I contribute to CumulusCI through REST API file synchronization tasks for enterprise artifact management.
