About
Most of what I enjoy is uncomfortable at first: heavy things at the gym, long climbs, airports at five in the morning, the occasional essay I didn't plan on writing. Software fits the pattern — the good problems push back for a while before they give. I'm a software engineer: backend services, data pipelines, and the infrastructure holding them up. Currently at Rippling, before that Morgan Stanley and ButterflyMX.
Most of it has been plumbing: streaming pipelines, event-driven services, and the slow business of working out why something gets slow every third day. Some of those systems had a model in them, which changed the failure modes more than it changed the job. What holds under load, what a dependency does when it starts lying to you, how you know the thing actually works — those questions stayed the same.
I like systems that still make sense once I've stopped touching them.
Experience
Aug 2026 — Present Software Engineer II, Backend · Rippling
Backend on payroll onboarding. Currently in the read-it-before-you-touch-it phase, which is the part I like anyway.
- Backend
- Distributed Systems
May 2026 — Aug 2026 Software Engineer, AI Platform · ButterflyMX
Internal platform for building AI agents and running them where people work.
- ▹Built the eval system. Most of the work was settling what “better” meant; the code was the easier half.
- ▹Before it, checking whether an agent change helped meant opening a chat and reading the reply. It runs before publish now.
- ▹Also the retrieval path, and getting one agent to behave the same in chat, Slack, a widget, and over the phone.
- TypeScript
- GraphQL
- Evaluation
- RAG
Nov 2023 — May 2026 Software Engineer II · Morgan Stanley
Backend and real-time systems in a regulated environment, and the first wave of LLM work that ran on top of them.
- ▹Took an annotation platform from 10K to 100K+ tasks a day by moving it off a synchronous monolith onto Kafka. P95 dropped from 800ms to 150ms.
- ▹Spent four weeks chasing a latency spike that showed up every few days. It was a Redis eviction caused by an unrelated batch job.
- ▹Built retrieval for fraud analysts that could only answer from evidence it had retrieved, and sent anything under the confidence threshold to a person.
- ▹Rebuilt the evaluation after a model beat every benchmark and landed worse. Benchmarks are a proxy, and proxies drift.
- ▹Two systems from this stretch are now granted patents.
- Python
- Kafka
- Flink
- AWS
- Terraform
Feb 2023 — Nov 2023 Software Engineer I · Morgan Stanley
Backend work, and some of the earliest production LLM deployments there — a retrieval-backed pipeline for modernizing legacy code, and the human feedback pipeline behind it.
- Python
- RAG
- RLHF
- AWS
Jul 2021 — Sep 2022 Regional Associate · Accelerator Intern · Hult Prize Foundation
Backend for a global competition platform, and an early lesson in the distance between a load test and what people actually do at peak.
- Node.js
- PostgreSQL
- Distributed Systems
Projects
Citation-Grounded Knowledge Agent
A multi-step agent built around one rule: every answer traces back to a source, enforced at generation rather than checked afterward. If retrieval isn't confident enough to support a grounded answer, it returns nothing.
- ▹Routes each query to the cheapest model that clears the quality bar, so inference cost stays a design consideration instead of a surprise.
- LangGraph
- AWS Bedrock
- pgvector
- LangSmith
- Python
Aarogya — Privacy-Preserving Mental Health Risk Detection
An NLP pipeline for early detection of depression and suicide-risk signals, with privacy treated as an architectural constraint. The hard part was deciding what a detection system that respects the person it's reading actually looks like.
- Python
- NLP
- Privacy-Preserving ML
- AWS
Complaint Routing via Textual Analysis
A text-classification pipeline that routes public and financial complaints to the right department, built through comparative experiments rather than a single model choice.
- Python
- Machine Learning
- Text Classification
Stack & Tools
Backend & Infrastructure
- Python
- TypeScript
- Node.js
- GraphQL
- REST APIs
- PostgreSQL
- pgvector
- MongoDB
- Kafka
- Flink
- AWS
- Terraform
- Docker
- CI/CD
AI & Agent Systems
- LangGraph
- LangChain
- AWS Bedrock
- RAG & Retrieval
- LLM Evaluation
- Agent Orchestration
- RLHF & Preference Data
- Confidence Routing
- Structured Output Enforcement
Certifications
- AWS Certified Machine Learning — Specialty
Research
- 2022
Context-Enriched Machine Learning-Based Approach for Sentiment Analysis
Research Publication · Apr 2022
- 2020
Comprehensive Review of Text-Mining Applications in Finance
Q1 Journal · Nov 2020
- 2020
Interplay of Machine Learning and Software Engineering for Quality Estimations
Research Publication · Nov 2020
- 2020
BioUAV: Blockchain Framework for Digital Identification in Next-Gen UAVs
Research Publication · Sep 2020
- 2020
Comparative Study of Sentiment Analysis and Text Summarization for Commercial Social Networks
Research Publication · Jul 2020
Education
- Buffalo, NY
University at Buffalo, SUNY
M.S., Computer Science
- Ahmedabad, India
Nirma University
B.Tech, Computer Engineering