I am a software engineer with four years of experience building production
systems at scale. My work sits where distributed systems meet data: transactional
services that hold up under concurrent load, event-driven pipelines that process
streams in real time, and, most recently, sensor ingestion and signal processing
for industrial vibration data. I hold an M.S. in Data Science from Indiana
University and write mostly in C++, Java, and Python.
1Experience
Present
Reed Family Companies
Software Developer
Designed a data ingestion pipeline that collects vibration data from rollers over MQTT with 99.998% accuracy.
Developed an FFT-based algorithm to extract the density signal from raw vibration data, and built a dashboard to monitor it.
Ithena
Full-Stack Software Developer
Designed a full-stack web application that builds dashboards and generates over 10 KPIs from user files or databases.
Architected the backend on LangChain and ChromaDB for vector embedding storage and retrieval, supporting up to 1 million concurrent conversations.
Reworked the chat interface with Material UI and message streaming, reducing render time by 15%.
Implemented authentication with Auth0 and per-user API rate limiting to protect against abuse while keeping performance consistent under high traffic.
Innomountain Inc.
Software Developer
Designed and implemented a low-latency backend for a parking management system, monetizing over 1,000 workplaces and malls in Massachusetts.
Engineered slot-booking APIs with transactional locking for safe concurrent reservations at peak traffic above 10,000 requests per second.
Automated backend infrastructure with AWS CloudFormation and EKS, cutting infrastructure cost by $4,000 per month.
Reduced database query latency by 30% through indexing strategies and query optimization.
Tata Consultancy Services
Software Engineer
Developed backend REST APIs in Spring Boot and Java for Citibank’s Institutional Clients Group, enforcing strong consistency across distributed services and improving transaction processing time by 21%.
Integrated Apache Kafka for reliable batch processing of over 50,000 transactions.
Automated CI/CD pipelines, reducing manual effort and deployment errors by 40%.
Orchestrated containerized deployments on Kubernetes with replica sets and health checks, sustaining 99.9% uptime.
2Open source
Transparent CSRF token recovery in Apache Superset
Expired CSRF tokens caused JSON API mutations to fail and discarded the user’s
UI state. The fix makes the backend error middleware emit structured CSRF error
responses, and adds a thread-safe retry in SupersetClient that
refreshes the token and retries the request once, without a page reload.
Concurrent failing requests share a single refresh promise, which prevents a
stampede of token refreshes.
A multi-threaded key-value store built from scratch, with a thread-per-connection
model and synchronization primitives guarding shared state. It implements the
Redis Serialization Protocol (RESP), including pipelining, so the official
redis-cli and standard client libraries can talk to it. Core
structures include a thread-safe hash table with incremental resizing and skip
lists for sorted sets, giving O(1) and O(log N) operations
respectively. Durability comes from a background append-only file (AOF) that
stays off the main execution thread.
A notification service that sends a low-latency WhatsApp alert whenever an
aircraft passes over a subscriber’s city. User locations are matched to flights
with bounding-box filtering and H3 spatial hashing, which cut notification
computation time by 33% and scales to thousands of simultaneous users. A
Kafka-based, event-driven pipeline orchestrates flight data ingestion,
user–flight matching, and alert generation over streaming updates.