Tech Lead for Ramp's Fraud detection platform

• Optimizing IP and MPLS routing protocols (BGP, OSPF, IS-IS, PIM, LDP, RSVP) and Layer 2 protocols (ARP, STP)

• Researching tensor and model parallelism for distributed LLM training to reduce network bottlenecks

• Engineered an end to end no-code rules engine, involving a drag and drop UI, an expression compiler/interpreter, and a python code emitter to enable fraud analysts to develop advanced vectorized DSL fraud rules without prior knowledge

• Researched streaming processors + parallelization for time travel queries to build a scalable machine learning feature store

• Lectured 150+ engineers on LLM powered structured extraction and built an internal tool to optimize prompt parameters

• Part of a small team building fintech products that drive 40% of overall company revenue

• Interfacing with Operations and Product team stakeholders to continue to support last year's intern project

• Upgraded entire API platform for handling of PCI-DSS sensitive info and enabled new processors (Marqeta, Adyen, Stripe)

• Owned an UX and code for an endpoint used to provide thousands of customers with new virtual credit cards

• Scribed architecture discussions and groomed tickets on a backend system handling >10 billion USD of TPV per month