Software engineer building full-stack systems that hold up in production, from React interfaces and API contracts through CI/CD and cloud deployment. This past year I've been working on AI agent infrastructure. Based in San Francisco, CA.
I'm a software engineer at PayPal, working on the internal Agent Toolkit. It decides which checkout, merchant, and invoicing operations an AI agent is allowed to perform, and holds that line when one asks for more.
Most of what I build has to hold up when something goes wrong. A wallet ledger that rebuilds every balance by replaying its event log. A proxy that cuts token spend and then has to prove it didn't change the answer. I like working under that kind of constraint. It makes you say out loud what happens on failure instead of hoping it doesn't come up.
Before this I spent two years at Deloitte on an internal search platform, getting compliance and policy documents in front of the people who needed them without getting them in front of anyone else. Different domain, same shape of problem: someone eventually asks what the system did, and the honest answer has to already be written down.
The agent work turned out to be made of familiar parts: keeping events in order across processes, checking that a system actually did what it said, and making any of it readable to the person watching. What's new is the blast radius, since the tool on the other end moves money.
I also taught Python and AI at Fullerton for a year and a half, around 40 students a section. Reading that many submissions changes how you write code. You start noticing every way an instruction can be misread.
I'm in San Francisco, CA. If you want to talk about backend systems, distributed systems, or agent infrastructure, email me.
Aug 2024 to May 2026
California State University, Fullerton
Fullerton, CA
Aug 2020 to May 2024
Atharva College of Engineering, University of Mumbai
Mumbai, India
Sep 2025 to Present
PayPal / California, USA
Building the Java and Python services behind an internal Agent Toolkit that gives AI agents scoped, audited access to PayPal's Checkout, Merchant, and Invoicing APIs. The schema-driven module design cut the work of adding a new agent capability by about 25%.
Jan 2025 to May 2026
California State University, Fullerton / Fullerton, CA
Taught Python Programming (CPSC 223P) and Artificial Intelligence (CPSC 481) to sections of 35 to 40 students, covering data structures, algorithms, object-oriented design, and neural networks.
May 2022 to Jun 2024
Deloitte / Mumbai, India
Built an internal knowledge-search platform on Python and FastAPI that let employees find compliance, policy, and operational documents across roughly 1.5 TB of enterprise content, serving over 5,500 searches a day at under three seconds each.
2026
Agent infrastructure
1st place, Mistral Hackathon ($3,000)
A repository rendered as a city you can walk, where every building is a real system and every grade traces back to evidence.
2026
Agent infrastructure
A local proxy that makes AI agents cheaper to run, with a cost ledger that proves it.
2026
Backend systems
An event-sourced wallet ledger where the audit trail and the data are the same artifact.
2026
Agent infrastructure
Five agents you watch work instead of reading a transcript afterward.
2026
Backend systems
A serverless document pipeline that turns scanned paperwork into queryable data.
2026
Agent infrastructure
Coding agents that verify their own work in a real browser, driven from Discord, Slack, or voice.
2026
Agent infrastructure
Agents that decide what intervention to run by reading what worked last time.
Object-oriented design, data structures and algorithms, multithreading and concurrency, shell scripting.
REST API design and versioning, SOAP, WebSockets, OAuth 2.0, JWT, microservices, event-driven architecture, CQRS, transactional outbox, saga orchestration, streaming proxies.
High- and low-level design, scalability, high availability, fault tolerance, service discovery, CAP theorem, SOLID principles, software design patterns.
DynamoDB, schema design, migrations, indexing, query optimization, transactions, stored procedures, data modeling, append-only event stores.
Lambda, EC2, S3, SQS, RDS and Aurora, IAM, VPC, CloudFormation, CloudWatch, Cognito, auto scaling. Azure Blob Storage, Cognitive Search, Active Directory. BigQuery and Vertex AI. Infrastructure as code.
Azure DevOps, ELK stack, build and deployment automation, containerization, structured logging and distributed tracing.
LLM gateway proxying, agent orchestration and multi-agent workflows, RAG, embeddings and vector search, fine-tuning, prompt engineering, context optimization, skill compilation and replay, output parity grading. Transformers and NLP, ML pipelines, model evaluation and deployment.
Component architecture, state management and hooks, responsive design, real-time data rendering, frontend performance.
Mockito, unittest, Newman, SonarQube. Unit, integration, contract, and regression testing, test-driven development, end-to-end and performance testing. Peer code review, Agile and Scrum, technical documentation.