Khushal Jethava
About
I'm a Machine Learning Engineer with 4+ years of experience designing and shipping production AI systems across document intelligence, computer vision, and generative AI. I've built Document AI products with fine-tuned transformers, room-transformation pipelines with ControlNet, car damage detection with Detectron2, and generative chatbots powered by LSTM and BART.
My current focus is advanced GenAI work — LLM applications, RAG systems, AI agents, and image-to-avatar transformation — and I write about it all on my blog, where I share practical AI and Python tutorials with hands-on code.
Skills & Technologies
Experience
Artificial Intelligence Engineer — Codiste
- Building advanced GenAI solutions including deepfake technology and image-to-avatar transformation pipelines.
- Developing LLM-powered applications, AI agents, and production RAG systems.
Machine Learning Engineer — Maknon Private Ltd.
- Performed image segmentation to identify and categorize floors, walls, and objects; fine-tuned models for accuracy.
- Implemented room transformations with ControlNet and built backend APIs to serve AI functionality.
Artificial Intelligence Engineer — SoluSoft Corporation
- Developed a Document AI product for classification and information extraction using fine-tuned transformers with Django and FastAPI.
- Built barcode detection (YOLO, OpenCV) and a car damage detection system (Detectron2, CNNs).
- Prototyped generative chatbots with LSTM sentiment analysis and BART dialogue summarization; led the project team.
Highlighted Work
Document AI Platform
Document classification and information extraction product built on fine-tuned transformers, served with Django and FastAPI.
Room Transformation with ControlNet
Image segmentation of interior spaces with high-quality generative room redesigns driven by ControlNet.
Car Damage Detection
Automated vehicle damage assessment framework using Detectron2 and convolutional neural networks.
Barcode Detection System
High-throughput barcode detection in documents using YOLO and OpenCV for efficient processing.
Generative Chatbots
Conversational AI with LSTM-based sentiment analysis and BART-based dialogue summarization.
Image-to-Avatar GenAI
State-of-the-art generative pipelines for deepfake technology and image-to-avatar transformation.
Latest from the Blog
Jun 27, 2026Build a Reasoning LLM from Scratch in Python
Build a reasoning LLM from scratch in Python — a BPE tokenizer, RoPE attention, SwiGLU transformer blocks, and chain-of-thought dual-loss...
Jun 23, 2026Understanding Embeddings: From Word2Vec to Modern LLMs
Learn how embeddings work in Python, from Word2Vec to modern transformer-based embedding models. Covers vector arithmetic, cosine similar...
Jun 19, 2026 View all postsGet in Touch
Open to collaborating on Generative AI, LLM, and computer vision projects. The fastest way to reach me is by email or LinkedIn.