01-Software
Software projects
Most of my software projects lean backend. I usually end up enjoying the parts that deal with systems behavior, infrastructure, databases, or tools that do something concrete.
02-Languages
Python / Java / C / C++ / TypeScript / JavaScript / SQL / Shell
03-Frameworks
FastAPI / Spring Boot / React / Next.js / Django / Flask / Node.js
04-Infrastructure
Docker / Kubernetes / Terraform / Ansible / AWS / Jenkins / Linux / Kafka
05-Project
c_toralizer
Minimal preload-based networking shim that transparently routes TCP traffic through a SOCKS4 proxy by intercepting connect() at runtime.
Tech stack
C / SOCKS4 / LD_PRELOAD / DYLD / Linux
- transparent TCP tunneling with runtime interception
- lightweight C implementation with low overhead
- works with many CLI tools where library injection is allowed
- clear boundary conditions documented for SOCKS4
06-Project
Recreating the System-Wide FD Tables
Systems utility for inspecting Linux file descriptor tables by scanning /proc once and generating per-process, system-wide, vnode, composite, summary, and threshold-based views from shared collected data.
Tech stack
C / Linux / /proc / POSIX / Make
- parses CLI flags for per-process, system-wide, vnode, composite, summary, threshold, and pid-scoped output
- collects open file descriptor records from /proc with pid, fd number, link target, and inode metadata
- stores results in a dynamic table so multiple reports can be generated from one scan
- modular C design split across parser, proc, table, output, and main execution modules
07-Project
PostgreSQL Locks & Partitions
Database systems lab reproducing lock-manager behavior on partitioned PostgreSQL schemas under concurrent load.
Tech stack
PostgreSQL / Python / Bash / SQLAlchemy
- benchmarked lock behavior across partition/index fan-out
- simulated high-concurrency query/update workloads
- built lock-inspection scripts over pg_locks and blockers
- documented production tradeoffs for schema partitioning
08-Project
Column vs Row-Oriented Databases
Comparative benchmark project evaluating OLTP/OLAP behavior across row-oriented and column-oriented storage engines.
Tech stack
PostgreSQL / DuckDB / Python / SQL
- 5M-row workload with repeatable benchmark scripts
- measured ingest, lookup, and aggregate query performance
- analyzed index behavior and storage-layout implications
- mapped findings to system design decisions
09-Project
AI-Powered Healthcare Platform (Hackathon)
Real-time triage backend with asynchronous event streaming for high-volume healthcare intake simulation.
Tech stack
Django / Kafka / PyTest / Docker
- Django REST APIs for triage workflows
- Kafka event streaming for async processing
- PyTest-backed reliability checks
- containerized deployment and CI integration
10-Project
Generative AI Voice Agent Application
Voice-agent platform with REST/WebSocket backends and deployment automation across cloud environments.
Tech stack
Next.js / FastAPI / AWS / Jenkins / Kubernetes
- FastAPI services with WebSocket support
- retrieval-augmented workflows and observability dashboards
- Jenkins CI/CD and AWS test environment provisioning
- reproducible containerized deployments
11-Project
Business Manager Network
Full-stack business platform with modular API architecture, analytics dashboards, and cloud deployment.
Tech stack
React / Django REST / PostgreSQL / AWS / Nginx
- modular React frontend + Django REST backend
- AWS EC2/Amplify + RDS deployment
- Nginx + Gunicorn production setup
- analytics dashboard and Route 53 DNS configuration
12-Project
Cloud & DevOps Infrastructure Automation
Infrastructure-as-code project automating secure cloud provisioning, deployment workflows, and observability.
Tech stack
Kubernetes / Terraform / Ansible / AWS / Prometheus / Grafana / Linux
- automated AWS provisioning with least-privilege IAM and VPC isolation
- Kubernetes deployment workflows with rollback-safe releases
- Prometheus/Grafana monitoring and alert visibility
- Git-driven CI/CD automation and server hardening
13-Project
Brain Tumor MRI Classification
Computer vision pipeline for classifying brain tumor MRI images into four categories.
Tech stack
PyTorch / CNN / Medical Imaging / Python
- CNN training with augmented MRI datasets
- structured batching and transform pipelines
- validation metrics and visual prediction overlays
- achieved 86% validation accuracy