#Software-Developer

DHRUV SEN

Welcome to my portfolio! I'm Dhruv Sen, a full-stack developer & ai engineer. I craft visually stunning, highly functional applications that deliver exceptional user experiences.

GitHub LinkedIn Email

About Me

My professional journey and technical expertise

I'm a passionate developer specializing in full-stack web and mobile development, machine learning, and AI-powered solutions. I love building products that blend cutting-edge technology with elegant user experiences. From crafting real-time AI assistants to designing NLP-powered detection systems, I bring ideas to life with clean code and thoughtful design.

India

LANGUAGES

Python

JavaScript

C++

Dart

STACK

MERN Stack

Next.js

React Native

Flutter

AI/ML

PyTorch

TensorFlow

Fine-Tuning (LoRA, PEFT)

RAG

LLMops

TOOLS

GitHub

Docker

AWS

Cloudflare

Linux

Education

My academic journey and learning milestones

CGPA: 8.2

B. Tech Computer Science & Engineering

Lovely Professional University

Percentage: 85.4%

Intermediate

St. Norbert School

Percentage: 87%

Matriculation

St. Norbert School

Experience

My professional journey and contributions

Great Chyren Welfare

Freelancer

May 2025 – Nov 2025

Designed an intuitive, user-centric UI/UX for an AI-powered wellness application, ensuring intuitive navigation and engaging user interactions.
Developed the application using Flutter, Firebase (with Authentication), and REST APIs, featuring both voice and text-based AI assistants, and improving response speed to 500 ms for real-time interactions.
Built a voice-enabled AI pipeline integrating speech-to-text (STT), RAG-based reasoning, and text-to-speech (TTS) for natural, context-aware voice assistance.

Selected Projects

A showcase of my recent work and technical expertise

Mar 2025

Dia Chat

On-Device AI Conversational Assistant

Fine-tuned Gemma-3 1B LLM with PyTorch and Unsloth to achieve 40 tokens/sec streaming responses, delivering Gen Z-aligned conversational experiences with authentic empathy.
Deployed full LLM inference directly in the browser using WebGPU and WebAssembly, ensuring zero-latency, privacy-first interactions without backend dependency.
Engineered a personality-driven AI that understands context, emotion, and nuance—built with Next.js frontend and Hugging Face model integration for seamless UX.

PyTorchHugging FaceNext.jsWebGPUWebAssemblyLLMFine-tuning

Dec 2025

GrievanceAI

Duplicate Grievance Detection

Architected an NLP-powered duplicate detection system for citizen grievance management using transformer embeddings (all-MiniLM-L6-v2, 384-dim) model and DBSCAN clustering, processing 100+ grievances per batch.
Orchestrated a multi-metric similarity pipeline (Cosine + Jaccard + N-gram) with adaptive threshold learning across 7 grievance categories.
Rolled out a microservice architecture with Flask embedding server, React dashboard, and Node.js backend featuring hierarchical deduplication and auto-categorization.

NLPTransformersFlaskReactNode.jsDBSCAN

Dec 2025

Animaex

Anime Director

Built a multi-agent manga generation system using LangGraph to orchestrate specialized agents (Script Writer, Storyboard Artist, Art Director), enabling structured and scalable creative workflows.
Developed real-time SVG manga panel generation using Gemini 2.5 Flash, eliminating traditional image generation bottlenecks and delivering lightweight, scalable vector artwork with dynamic styles (Shonen, Shojo, Seinen, etc.).
Implemented a high-performance streaming pipeline with FastAPI + WebSockets, integrated with Cloudinary CDN for instant panel delivery, and ChromaDB for persistent character/lore memory.