Software Engineer with a focus on AI/ML integration and full stack product engineering. I build systems that connect intelligent backends to clean, responsive frontends — from LLM-powered APIs to production-grade data pipelines.
My engineering philosophy is grounded in performance, scalability, and shipping with intent. During my internship at Incanto Dynamics, Bangalore, I worked as a Data Analysis & AI Intern, building real-world pipelines around the Gemini API and deploying models into live product workflows.
I approach every layer of the stack with the same rigour: architecting APIs that hold under concurrent load, designing MongoDB schemas that scale, and instrumenting React frontends that remain performant in production.
Core Engineering Values:
- Systems-first thinking over feature-first execution
- Measurable performance over abstract quality
- Composable architecture over monolithic solutions
- Continuous delivery over waterfall perfectionism
Open To:
Software Engineering · AI / ML Engineering · Full Stack Development · Backend Engineering · Data Engineering · AI Product Roles
| Domain | Proficiency | Details |
|---|---|---|
| LLM Integration | ████████░░ Advanced | Gemini API, prompt engineering, multi-turn conversation design |
| Generative AI | ███████░░░ Proficient | Image generation pipelines, AI-powered styling systems |
| Data Analysis | ████████░░ Advanced | Pandas, NumPy, exploratory analysis, visualization |
| Machine Learning | ██████░░░░ Proficient | Scikit-learn, feature engineering, model deployment |
| NLP & RAG | █████░░░░░ Developing | Retrieval-augmented generation, document QA |
| AI Product Engineering | ████████░░ Advanced | End-to-end AI feature integration in production stacks |
| Fraud Detection Systems | ██████░░░░ Proficient | Real-time stream processing, Kafka-based anomaly detection |
⬡ F1 Telemetry & Strategy AI Dashboard
An enterprise-grade Formula 1 telemetry analysis platform powered by the Gemini API and Scikit-learn. Processes live race telemetry data to surface real-time strategic insights, tyre degradation models, and lap-time prediction — built to handle the scale and latency demands of motorsport data.
| Attribute | Detail |
|---|---|
| Stack | Python · FastAPI · Gemini API · Scikit-learn · React · PostgreSQL |
| Scale | 500+ concurrent requests sustained under load |
| Performance | Sub-200ms API response at p95 under race-day simulation |
| AI Layer | Gemini API for natural language strategy summaries; Scikit-learn for tyre degradation regression |
| Impact | Full race-weekend simulation coverage — qualifying, pit windows, safety car scenarios |
| Repository |  |
The core engineering challenge was building a data ingestion layer fast enough to keep pace with telemetry streams while running ML inference and LLM summarization in parallel without blocking the request thread. Solved via async FastAPI workers with a Redis-backed job queue for model inference tasks.
⬡ AI-Powered Skill Bartering Platform
A full stack peer-to-peer skill exchange marketplace where users trade expertise instead of currency. Built on a MERN stack with Redis-backed session management and a Docker-containerised architecture designed to sustain high-concurrency matchmaking and real-time chat.
| Attribute | Detail |
|---|---|
| Stack | MongoDB · Express · React · Node.js · Redis · Docker |
| Scale | Architected for 1,000+ concurrent user connections |
| Performance | Redis caching reduces DB read latency by ~65% on hot skill queries |
| Security | JWT-based auth · Role-based access control · Input sanitization |
| Impact | Zero-cost skill acquisition model — replaces transactional friction with value exchange |
| Repository | |
The matchmaking engine uses a weighted compatibility graph built from user skill profiles and availability metadata, with Redis Sorted Sets serving ranked match results at constant-time lookup. Docker Compose orchestrates the full service mesh locally; the same config targets production with minimal delta.
⬡ AI-Powered Virtual Try-On & Stylist
A generative AI application that enables users to virtually try on clothing items and receive personalised style recommendations. Combines computer vision with a generative model backend to produce realistic garment overlays without physical product samples.
| Attribute | Detail |
|---|---|
| Stack | Python · Generative AI · FastAPI · React |
| AI Layer | Generative image synthesis · Style preference modelling |
| Performance | Inference pipeline optimised for sub-3s try-on render |
| Impact | Reduces purchase hesitation — bridges the gap between e-commerce browsing and fitting-room confidence |
| Use Case | Fashion retail, wardrobe planning, personal styling tools |
| Repository | |
The primary engineering constraint was balancing image quality against inference latency. Implemented a two-stage pipeline: fast low-res preview generation in under 800ms, with a high-res synthesis pass queued asynchronously and delivered via polling.
Bangalore, India · 2026
Worked within an applied AI product team, contributing to data analysis workflows and AI feature integration. Built and shipped production-grade pipelines using the Gemini API, transforming raw data into structured, actionable model inputs.
- Designed and deployed LLM-backed analysis pipelines using the Gemini API in a live product environment
- Built data preprocessing and feature engineering workflows in Python, improving downstream model reliability
- Contributed to AI feature integration across FastAPI-based backend services
- Conducted exploratory data analysis across business datasets, surfacing signals for product and engineering decisions
- Collaborated with engineering leads on architecture decisions for AI module scaling
| Recognition | Details |
|---|---|
| 🏗️ Production Deployment | Shipped live TCS NQT prep tool to Vercel with 136 curated problems — real users, real usage |
| 🤖 AI Product Engineering | Built and deployed LLM pipelines at Incanto Dynamics using Gemini API in production |
| 📊 Internship: Data Analysis & AI | Industry experience at Incanto Dynamics, Bangalore — AI feature integration & data pipelines |
| 🧩 DSA Tracker: 42/57 PYQs | Completed 73%+ of a self-curated 57-question TCS NQT PYQ tracker (32/33 Easy · 10/16 Medium) |
| 🧱 Concurrent Systems Design | Architected system handling 1,000+ concurrent connections with Redis + Docker |
| 📈 F1 Telemetry at Scale | Built AI dashboard sustaining 500+ concurrent requests with sub-200ms p95 latency |
| 📝 LinkedIn Technical Writing | Weekly posts on Docker internals, LeetCode DP patterns — consistent technical knowledge sharing |
| 🎯 Full Stack × AI Intersection | Projects span the full product surface: ML inference → API → frontend → deployment |
learning:
- AWS Solutions Architect (SAA-C03) — post CLF-C02 certification path
- Advanced DSA: Dynamic Programming, Graphs, Trees (Striver + Aditya Verma)
- System Design fundamentals for SDE interviews at scale companies
- Azure AI Fundamentals (AI-900) via Microsoft AI Skills Yatra
building:
- Expanding F1 Telemetry Dashboard with real-time Gemini API strategy summaries
- Hardening Skill Bartering Platform for production-scale concurrency
- Technical content pipeline for @thequietdev (tech · finance · career)
exploring:
- Open source contributions to LangChain / LlamaIndex
- Kafka-based real-time fraud detection architecture
- RAG pipeline design patterns for enterprise search
open_to:
role_types:
- Software Engineer (Full Stack / Backend)
- AI / ML Engineer
- Data Engineer
- AI Product Engineer
minimum_ctc: 5 LPA
location: Bangalore (on-site / hybrid preferred) | Remote considered
notice_period: Immediate
preferred_stacks:
- Python + FastAPI + LLM stack
- MERN + Redis + Docker
- Data pipelines + AI feature integration
"The best engineers don't just write code — they build systems that outlive the sprint."