Diagnosed and resolved MongoDB performance issues including
over-fetching (1M+ unnecessary documents), N+1 query patterns, and
excessive frontend API calls, significantly reducing data transfer costs
and eliminating production timeouts; coincided with ~30% infrastructure
cost reduction
Improved system performance by refactoring schema anti-patterns and
enforcing MongoDB best practices
Enhanced observability across microservices by implementing
OpenTelemetry, enabling real-time latency monitoring and correlated
logging
Maintained and debugged Builder, a proprietary DSL compiler critical
to data accuracy, enabling nutrition experts to author 1,000+ data
transformation formulas compiled to JavaScript at runtime; improved
developer experience by fixing error propagation and supporting
hot-reloading for production systems processing 1M+ products
Conducted comprehensive code reviews for 9-person engineering team,
identifying MongoDB anti-patterns, data over-fetching issues, and
performance bottlenecks; frequently served as primary reviewer during
high-velocity delivery periods, ensuring architectural best practices
and preventing production performance issues
Facilitated post-acquisition integration of 7-person Product
Intelligence team by developing automated documentation system (Mongoose
schemas to Markdown with cross-references), establishing development
environments, and providing ongoing architectural guidance for feature
implementation
Developed full-stack image processing pipeline managing 100K+ images
with priority queuing for batches up to 10K images; implemented S3
streaming, automated zip/folder organization, and identified race
conditions, implementing MongoDB document locking to prevent concurrent
modification during downloads
Built self-service script execution platform enabling operations
team to run data processing and maintenance scripts via web interface
with real-time logging; implemented local development environment with
temporary BigQuery tables, eliminating UAT deployment requirements for
testing
Deployed and maintained production services on GCP (Cloud Run, Cloud
Functions, BigQuery) using Terraform and Concourse; developed React/MobX
admin tools and managed AWS S3 integrations for data storage
Pinto ~(Acquired by SPINS)
Software Engineer
~07/2021 – 03/2022
Proposed and built a static analysis tool to transpile Mongoose
schemas into Markdown documents for automatic domain documentation
Developed real-time retailer data quality scoring system serving
100+ retailers; implemented middleware with tree traversal algorithms to
merge grading feedback into retailer documents upon product data
updates
Built APIs and provided technical support for data engineering
workflows; wrote documentation and conducted code reviews for
Python-based data collection scripts, and developed ad-hoc queries for
operations team
Wakefern Food Corporation
Software Engineer
~05/2017 – 07/2021
Built full-stack project management system serving 6 project
managers and 100+ department staff; implemented JIRA proxy with Vue.js
frontend and PostgreSQL backend to enforce PM-only project state
controls and provide executive-level reporting with data rollups for
department director
Created an ETL pipeline in Python to ingest and clean IBM Weather
API data
Developed custom JavaScript SPA framework with routing, state
management, and component system using regex-based template parsing;
implemented custom directive syntax ($$for loops, dynamic replacements)
for internal applications in restricted open-source environment
Implemented shop-from-home feature across 100+ Shoprite stores using
geofencing technology; developed cross-platform location detection with
unified iOS/Android permission handling and state management for
automatic store notification upon customer arrival
Parsed and transformed XML
transaction data to DB2 for customer shopping trend analysis
Personal Projects
Algorithmic Trading System ~(Go, Python,
MongoDB)
Event-driven low-latency trading engine built with components like
Order Manager and Execution Manager. Uses Go channels to implement
Actor-like model to minimize contention. Includes custom REST/WebSocket
clients, supports 80+ active orders and 8 concurrent markets, and
features execution algorithms like TWAP.
A mail server fully compliant with the POP3 protocol. It uses
MailDir as the underlying storage layout, Sled to store users, and
Argon2 to secure passwords.
A simple Rust program to help solve the Sporcle quiz “US City
Triplets” utilizing web scraping to retrieve required inputs and file
parsing/scanning to identify potential solutions based on those
inputs