Selected work

Systems, AI, and product engineering

Five projects that show distributed systems depth, GPU/algorithms work, production AI, full-stack product delivery, and software that talks to hardware.

Systems

Distributed Event Processing Platform

18K events/secp95 < 120ms

Event-driven backend for high-throughput ingestion, persistence, caching, and horizontal scale across containerized services.

JavaSpring BootKafkaPostgreSQLRedisDockerKubernetes

View on GitHub →

GPU

CUDA Spatial Query System

100K+ q/ssub-ms latency

GPU-accelerated spatial query engine with R-Tree / KD-Tree indexing and PostGIS-backed storage for high-throughput geometric workloads.

C++CUDAPostgreSQLPostGISR-TreeKD-Tree

View on GitHub →

AI Support Agent

1,000+ daily queries−35% tickets

Production RAG agent with retrieval, confidence gating, and escalation — not just a chat UI. Built in a professional role for a live product environment.

1User
2API
3Retriever → Postgres / vectors
4LLM
5Confidence
6Respond or escalate

PythonLangChainRAGPostgreSQLAWSReact

Built at work — source is proprietary; no public repository.

Product

PixelPen

Shipped to productionAI writing & Q&A

Full product surface: blogging, auth, AI-assisted writing/summarization, and personalized learning. Previously scaled to 10,000+ registered users before the service was wound down.

Next.jsReactPythonMongoDBPrisma

Open live app →

Hardware

Real-Time Sensor & Hardware Integration

Live WebSocket telemetryESP32 + simulator ingest

End-to-end pipeline from ESP32 environmental sensors (or a Python simulator) through a serial/MQTT gateway into a FastAPI backend, with a live React dashboard for monitoring and threshold alerts.

C++PythonFastAPIReactWebSocketMQTTDocker

View on GitHub →

I'm Aniket - a software engineer, blogger and tech enthusiast. Thanks for checking out my site!

General

Specifics

Connect