Akshay Kamath

Computer Science @ SJSU · Software Engineer

Building production systems across backend, full-stack, and AI/ML.

GitHubResume
Hero photo 1

Experience

  • Developed a LangChain/GPT-4 agent flagging $100M+ subscription variances; cut analyst review time by 85%
  • Made a Streamlit dashboard to analyze subscription + backlog records and surface variance drivers to analysts
  • Coded Python/SQL pipelines to process over 200,000+ records across 11 tables; exposed metrics via REST APIs

Additional Details

  • Built LangChain agent with custom prompt engineering to identify subscription anomalies
  • Streamlit dashboard with interactive filters and drill-down capabilities
  • SQL optimization reduced query time from 45s to 3s on 200k+ record joins
  • Analyzed 12,000+ parking lots into 46 priority sites across 3 cities using k-means for charger deployment
  • Shipped Flask APIs + React/Folium map with PostgreSQL for self-serve owner lookup and ops reporting
  • Documented APIs, schemas, and configs to enable smooth handoff to engineering and operations teams

Additional Details

  • K-means clustering with feature engineering (traffic, parking density, accessibility)
  • Interactive Folium map with heatmaps and site markers
  • PostgreSQL geospatial queries for proximity analysis
  • Designed Java/Spring Boot microservices; Dockerized and deployed to Cisco CAE cloud via Jenkins CI/CD
  • Implemented Kafka pipelines for tax + revenue data (10,000+/day) and cut peak-load latency 25%
  • Instrumented services with metrics + alerts for observability to speed incident triage and meet internal reliability targets

Additional Details

  • Spring Boot microservices with RESTful APIs and service discovery
  • Kafka consumer groups for parallel processing of tax events
  • Prometheus + Grafana dashboards for real-time monitoring
  • Built TensorFlow battery-health models + tooling; improved MAE 15% on NASA and CALCE datasets
  • Optimized self-attention denoiser on 100,000+ cycles; cut prediction time 20% for monitoring
  • Wrote eval scripts, docs, and config templates so lab members can reproduce and extend experiments

Additional Details

  • Transformer-based denoising autoencoder for battery capacity prediction
  • Hyperparameter tuning with Optuna for model optimization
  • Reproducible experiment pipeline with MLflow tracking

Projects

Sort:

Homeplate

DoorDash-style meal-share platform

A full-stack meal-sharing app connecting home cooks with diners. Features real-time booking, payments, reviews, and a Dockerized development environment.

Next.jsReactTypeScriptNode.jsExpressPostgreSQLStripeDocker
  • Designed relational schema + REST APIs for cook discovery, booking flows, and reviews
  • Integrated Stripe checkout for payments and built end-to-end booking lifecycle
  • Dockerized services for reproducible local setup

Strands

Threads-like social network

A social media platform inspired by Threads, featuring posts, follows, personalized feeds, and real-time search. Built with Spring Boot and React.

Spring BootReactPostgreSQLDocker
  • Built core social features (posting, following, feed, search) with clean full-stack architecture
  • Dockerized entire stack and automated schema + seed SQL for repeatable dev environment
  • Designed API contracts to keep frontend iteration fast and backend changes safe

File Management System

FAT-style virtual filesystem

A virtual file system CLI implementing FAT allocation with persistent metadata, free-space tracking, and an open-file table. Supports core operations like mkdir, create, open, read, write, mv, and rm.

Python
  • Implemented FAT/free-map allocation with persistent metadata for crash-resilient state
  • Built core commands (mkdir/create/open/read/write/mv/rm) with validation and edge-case handling
  • Wrote comprehensive tests to catch boundary conditions and consistency errors

Now

What I'm currently focused on

Applying to Summer 2026 internships

Refining skills in backend and full-stack development

Building out startup ideas

Building knowledge in the quant dev domain

Last updated: January 2026

Contact

akshaykamath2004@gmail.com

Drop a quick note here.