Built a production-quality Spider-Verse style video LoRA for Wan 2.1 T2V 1.3B. Engineered a dataset pipeline for 1,239 movie clips (8,969 extracted frames), including caption generation and dataset cleaning/validation. Fine-tuned using mixed-precision FP16 training, AdamW8Bit optimizer, FlowMatch scheduler, and gradient checkpointing on an RTX 3050 Laptop (6 GB VRAM). Integrated GGUF-quantized Wan models into ComfyUI for efficient local prompt-controlled text-to-video inference with GPU optimization.

AI Model Management Platform

January 1, 2026 - Present

Developed a Spring Boot backend to manage AI models, datasets, training jobs, and inference pipelines. Designed REST APIs for model registration, versioning, checkpoint management, and experiment tracking. Integrated Python-based Stable Diffusion, Wan 2.1, and local RAG services via REST APIs for AI workflow orchestration and added metadata management for datasets, outputs, and evaluation.

RAG Chatbot using Qwen 3B + FAISS

January 1, 2026 - Present

Built a fully local Retrieval-Augmented Generation chatbot using Qwen 3B for document-aware question answering. Implemented PDF parsing with PyMuPDF, semantic chunking, metadata extraction, and dense vector indexing using FAISS. Designed a retrieval pipeline that injects top-k relevant context into a local LLM for grounded responses. Added a Streamlit UI for PDF upload, semantic search, and conversational querying with GPU-accelerated local inference (no cloud LLM APIs).

PixelLoRA Production Studio

July 1, 2025 - Present

Implemented a specialized pixel-art generation workflow using DreamShaper 8 with custom Pixel Art LoRA fine-tuning. Built the dataset pipeline: collected and processed 2,186 images, removed 164 duplicates, prepared 2,001 cleaned images with 2,001 captions. Ran LoRA training experiments with GPU/VRAM-aware optimization and created local inference workflows using Gradio and ComfyUI, including generation and output evaluation.

CARTOONIX – Prompt to Cartoon / Sketch Generator

July 1, 2025 - Present

Developed a multimodal diffusion pipeline for prompt-to-cartoon, sketch, pixel art, and comic-style image generation. Built a Stable Diffusion pipeline using DreamShaper v8 and trained custom LoRAs for style control. Curated and captioned 6,000+ images across 20+ styles. Added ControlNet and OpenCV edge guidance to improve sketch/structure consistency. Created Gradio-based desktop apps for text-to-image and image-to-image generation on an RTX 3050 Laptop.

Violence Detection using CNN + LSTM

December 1, 2024 - February 28, 2025

Developed a real-time violence detection system using a hybrid CNN + LSTM architecture for spatio-temporal video analysis. Trained and evaluated the model on RWF-2000 achieving 92% classification accuracy. Implemented OpenCV real-time inference with automated Telegram alerts for detected incidents and evaluated robustness under lighting variations, motion blur, and playback speed changes.

Violence Detection Engineer (CNN + LSTM)

December 1, 2024 - February 1, 2025

Built a real-time violence detection system using a hybrid CNN + LSTM architecture for spatio-temporal video analysis. Trained on RWF-2000 achieving 92% classification accuracy. Implemented OpenCV-based real-time inference with automated Telegram alerts and evaluated robustness under lighting variations, motion blur, and playback speed changes.

Artificial Intelligence Intern at Agratas Academy Pvt. Ltd. (MSME, Govt. of India)

July 1, 2024 - September 1, 2024

Completed industry-guided training in Artificial Intelligence and Machine Learning. Worked on AI model experimentation and development workflows and applied problem-solving during guided tasks. Gained exposure to real-world AI engineering and deployment practices.

AI & Data Science Engineer – Academic Projects at MES College of Engineering, Kuttippuram

October 1, 2022 - May 1, 2026

Developed deep learning models for computer vision, multimodal AI, and sequence learning using PyTorch and TensorFlow. Built end-to-end AI pipelines including dataset engineering, model training, inference, and evaluation. Developed diffusion model applications, LoRA fine-tuning workflows, video generation pipelines, and Retrieval-Augmented Generation systems.

Education

B.Tech – Artificial Intelligence and Data Science at MES College of Engineering, Kuttippuram

October 1, 2022 - May 1, 2026

Higher Secondary Education (Science) at APJ Abdul Kalam Technological University

January 1, 2020 - January 1, 2022

High School Diploma – Mechanical / Automobile at Government Technical High School, Kuttippuram

January 1, 2017 - January 1, 2020

B.Tech – Artificial Intelligence and Data Science at MES College of Engineering, Kuttippuram (APJ Abdul Kalam Technological University)

October 1, 2022 - May 31, 2026

Higher Secondary Education (Science) at GHSS Edappal (Higher Secondary Education Board)

January 1, 2020 - December 31, 2022

High School Diploma – Mechanical / Automobile at Government Technical High School, Kuttippuram (State Board)

January 1, 2017 - December 31, 2020