I'm a Software Engineer who ships thoughtful products from Mumbai to the world
I believe the best technology solves real-world problems seamlessly, with both elegance and simplicity.
My Motto
Thoughtful Code,
Meaningful Impact
My Capabilities
- Problem Solving
- Scalable System Design
- Rapid Prototyping
- User-Centric Solutions
- Clean Code Architecture
- Team Collaboration
- Continuous Learning
Experience
1.3+ Years
Users Reached
100K+
Core Developer
Nutriscan App
Projects Completed
20+
Professional Journey
A timeline of roles, responsibilities, and achievements across engineering.
- Led Flutter app development across iOS, Android and Web with clean architecture
- Built Python backend services and deployed to Google Cloud Run and Azure
- Built a RAG chatbot with text embeddings using Gemini/OpenAI; delivered additional AI features with prompt engineering
- Created cron jobs and Cloud Scheduler tasks for background processing and notifications
- Integrated payment gateway and implemented in-app subscriptions/purchases
- Designed Supabase PostgreSQL schemas and SQL functions; added WebSocket connections for realtime updates
- Led database migration for ~6,000 users ensuring data integrity and minimal downtime
- Contributed Flutter features and fixes with Provider state management
- Assisted Python service setup and CI/CD pipelines
- Integrated error monitoring (Crashlytics) and basic scheduling jobs
- Improved app performance by 60%+
- Helped prototype AI features and meal‑scanning flows
- Developed core gameplay flows and responsive UI in Flutter
- Implemented state management with Provider; reduced unnecessary rebuilds
- Built smooth, interrupt‑safe animations targeting 60 FPS on mid‑tier devices
- Added centralized error handling and graceful fallback UX
- Created reusable widgets and theming to speed up iteration
TechStack
Technologies and tools I work with to build innovative solutions.
AI / ML
Languages
Backend
Frameworks / Engines
Tools
Colab
Hugging FaceProjects & Innovations
A selection of featured work across AI, mobile, and cloud.
Fine-tuned Multimodal AI for Product Recognition
Fine-tuned Qwen2.5‑VL with Unsloth for practical vision‑language tasks (ID/name/count extraction) with a faster, parameter‑efficient training recipe; accompanied by a full walkthrough.
Health Risk Prediction with NHANES Data
Multi-output health risk prediction on the NHANES dataset (33K+ records). Trained Random Forest models with extensive feature engineering and ROC analysis; average AUC ≈ 0.85 with 0.90+ for diabetes and 0.86+ for hypertension.
Named Entity Recognition – OntoNotes 5 Dataset
Span-level NER model fine-tuned on the OntoNotes 5 dataset; published on Hugging Face. Supports common entity types (PERSON, ORG, GPE, LOC, DATE, TIME, MONEY) and suitable for production extraction workflows.
Smart Bill Book - AI Billing System
Implemented the Smart Bill Book application end‑to‑end in Flutter with AI‑powered bill extraction, robust billing flows, and production readiness.
AR Furniture App
Augmented Reality furniture visualizer built in Flutter: place, scale, and rotate multiple models in real spaces with stable tracking and lighting; optimized for smooth performance on mid‑tier devices.
Toon Hunter – Unity 3D
Mobile 3D action game built in Unity with physics‑driven gameplay, touch controls, and level progression; optimized for smooth performance on mid‑tier devices.
Technical Writing
Selected publications and articles on AI, mobile, and product engineering.
Fine-tuning Qwen2.5‑VL with Unsloth: A Complete Guide
Step‑by‑step guide covering dataset prep, PEFT setup, training/finetuning, and evaluation for vision‑language tasks.
ReadPersonalized Chatbot with LlamaIndex
Designing a retrieval-augmented pipeline for tailored conversational agents.
ReadIn-App Purchase Implementation
Implementing subscriptions with best practices for reliability and UX.
ReadGet inTouch
Have an idea or opportunity? Send a note or use a quick link.