About
Currently a Software & AI Engineering Intern at Openbots AI ↗︎, building HIPAA-compliant AI systems for LTAC hospitals. The team designs agents and ETL pipelines that compress admissions and clinical review from days down to hours.
Alongside the internship, researching with Dr. Chen Chen ↗︎ at FIU on conversation-aware AI agents for dementia memory recall. The work centers on a memory architecture designed so patients feel like they are talking to someone who genuinely knows them.
Personal arc: started undergrad at DJ Sanghvi College of Engineering ↗︎ in Mumbai, dropped out after first year and moved to the US to be inside the AI revolution rather than read about it from across the world. Iron Man got me into building, and that part hasn't changed.
Experience
Building HIPAA-compliant AI systems for LTAC hospitals. Optimizing production ETL pipelines in Python and .NET for large-scale patient records, and developing LangChain + FastAPI agents that automate multi-step clinical review. Designing RPA workflows integrated with hospital data systems that compress days of manual processing into hours.
Researching conversation-aware AI agents for dementia memory recall. Designing a memory architecture so patients feel like they're talking to someone who genuinely knows them. Ember is the front-end testbed for this work.
Manage the open-source contribution lifecycle for 50+ developers in CIS 3590. Enforce industry-standard Git workflows and review practices, and collaborate with the BoxLang engineering team to source and triage live open-source tickets students ship to production.
Architected a distributed networked state machine in Unity/C# with the Meta SDK, hitting 50ms end-to-end multiplayer latency. Cut rendering overhead 20% on passthrough-compatible interfaces through URP pipeline and Canvas Renderer tuning.
Optimized automated data processing pipelines for 50,000+ RFID entries with Python, SQL, and Kafka, cutting discrepancies 15%. Designed a Neo4j graph schema for device relationships and dependency mapping, enabling 30% faster fault isolation.