I'm a Full Stack + AI Systems Engineer specializing in high-performance backend architecture, distributed data systems, and production-grade AI/ML integration. I build systems designed to operate at scale — from LLM-powered applications to Kafka-driven pipelines processing billions of records.

My work spans the full engineering stack: secure REST APIs, real-time event systems, cloud-native infrastructure, and intelligent document processing. I contribute to open source and maintain published PyPI packages actively used in production.

Currently building NCode — a privacy-first AI code editor with local (Ollama) + cloud LLM support, built with Rust, Tauri, React, and Python.

Backend Engineering      ████████████████████  Django · FastAPI · Microservices · REST APIs
AI / ML Systems          ███████████████████░  RAG · LLM Integration · BERT · XGBoost · DNN
Distributed Systems      ██████████████████░░  Kafka · RabbitMQ · WebSockets · Async Python
Database Engineering     ████████████████████  PostgreSQL · pgvector · SQLAlchemy · Redis
DevOps & Cloud           ██████████████████░░  Kubernetes · Docker · AWS · GCP · Terraform
Security & Auth          ███████████████████░  JWT · OAuth2 · CSRF · RBAC · Custom Tokens
Data Pipelines           ██████████████████░░  ETL · Web Crawling · PDF · NLP Extraction

Languages

Backend & API

Frontend

Databases & Search

AI / ML

Messaging & Real-Time

Cloud & Infrastructure

Security & Testing

A cross-platform desktop IDE combining a VS Code-like editing experience with a fully integrated AI assistant. Runs entirely offline with local LLMs via Ollama, or optionally connects to cloud providers. Code never leaves your machine unless you choose.

Production-ready modular authentication system — JWT, OAuth2, custom middleware, Alembic migrations, PostgreSQL.

Django extension for complete authenticated user activity logging — pluggable, customizable, and published on PyPI.

A secure, pluggable activity logging system for Django applications. Tracks every authenticated user action with full request/response lifecycle visibility and JWT-based identity handling.

pip install django-activitylog-jwt

# settings.py
INSTALLED_APPS    = [..., 'activitylog']
MIDDLEWARE        = [..., 'activitylog.middleware.ActivityLogMiddleware']

What it provides:

• Request/response lifecycle tracking across all authenticated endpoints
• Per-user activity audit logs with timestamps and metadata
• JWT-based user identity extraction without additional dependencies
• Configurable log storage, filtering, and retention per project

A selection of production systems demonstrating cross-domain engineering depth:

Models & Techniques

• RAG (Retrieval-Augmented Generation) — production pipelines over SQL and document stores
• BERT — semantic search, document classification, NLP feature extraction
• XGBoost — tabular prediction, anomaly detection, feature engineering
• Deep Neural Networks — regression and multi-class classification pipelines
• KNN — similarity-based recommendation systems
• pgvector — vector similarity search directly on PostgreSQL
• Hugging Face — model loading, inference, fine-tuning workflows
• LLM Integration — OpenAI, Anthropic, Groq, Ollama (local inference)

Data at Scale

• Designed systems handling millions to billions of records
• Cursor-based and keyset pagination for large dataset traversal
• Kafka-backed distributed scraping and ingestion pipelines
• ETL pipelines with preprocessing, validation, and structured loading

These aren't just tools — they're how I approach every system:

• Scalability by default — systems designed to grow, not retrofitted
• API-first design — clean contracts before implementation
• Security in layers — custom auth, CSRF protection, RBAC, activity auditing
• Async-first where it matters — asyncio, asyncpg, async SQLAlchemy
• Minimal external dependencies — build for control and longevity
• Cross-stack integration — React + FastAPI + ML + Kafka + Kubernetes as one cohesive system

Bachelor of Science — Magadh University, Bihar, India

_{"Code with purpose, scale with precision, and automate everything."}