Headline

Machine Learning Engineer | Applied AI & Computer Vision | Researching, Building & Deploying AI Systems

About

I’m a tech-driven innovator passionate about building AI-powered solutions that create real-world impact. My expertise spans machine learning, computer vision, and embedded AI, with experience developing end-to-end systems for healthcare, agriculture, and autonomous systems. I have hands-on experience with deep learning (CNNs, RKNN, YOLO), real-time inference on edge devices (NVIDIA Orin Nano and Rockchip RV1126), ONNX optimization, and scalable AI pipelines. My work focuses on integrating AI, AR, and embedded systems to design practical, impactful solutions. I’ve published research on Long-Term Temperature Forecasting using Machine Learning in the Mathematics (MDPI) journal and filed a patent for the SignToSpeech system, an AI-AR-based hardware that translates sign language to speech and vice versa. I’ve also earned top positions in national hackathons, including the FloodAI Hackathon at IIT Gandhinagar, and contributed to projects that apply AI for sign language translation, crop disease detection, and flood prediction using satellite data. My goal is to bridge AI and the physical world by building intelligent systems that improve lives. I’m always open to collaborations, innovative projects, and opportunities in Machine Learning Engineering, Embedded AI, and AR-driven solutions.

Location

Bengaluru

AI stack

claude code

claude

gemini

langfuse

langgraph

mcp

Work experience

Machine Learning at HyperNorm AI, Dec 2025 - Present · 10 mos, Bengaluru, Karnataka, India, Focused on building LLM-powered research agents and intelligent AI systems for financial research and stock analysis. Developing agentic workflows that combine LLMs, data pipelines, retrieval, and automated reasoning to turn complex financial information into actionable insights.

AI/Ml Engineer at Buze Platforms Private Limited, May 2025 - Dec 2025 · 8 mos, Ahmedabad, Gujarat, India, Worked in the field of computer vision and embedded software development. Key responsibilities and contributions: Designed and deployed real-time person detection, tracking, and face recognition systems using YOLO and InsightFace (Buffalo L), achieving 92% accuracy and 30 FPS on NVIDIA T4 GPU with TensorRT optimization. Developed C++-based face recognition modules for Rockchip RV1126 using MPP, RockX, MTCNN, and MobileFaceNet (via RKNN), enabling 12 FPS low-latency edge inference. Implemented bounding box–based distance estimation for real-time proximity analysis. Automated multi-platform deployment of CUDA, ONNX, and RKNN modules using CMake, supporting various embedded systems.

Research Intern at India Meteorological Department, May 2025 - Jul 2025 · 3 mos, Bhubaneswar, Odisha, India, Worked on a research-focused project titled "Temperature Prediction Using Time Series Forecasting" at the India Meteorological Department (IMD) under the guidance of Dr. Manorama Mohanty. Key responsibilities and contributions: • Collected and preprocessed large-scale meteorological datasets for temperature analysis. • Applied time series forecasting techniques (ARIMA, LSTM, etc.) for short-term and long-term temperature prediction. • Evaluated model performance using appropriate statistical metrics and improved prediction accuracy through iterative tuning. • Compiled findings and insights into a detailed research report for internal use and further study.

Student Researcher at Pandit Deendayal Energy University (PDEU), Oct 2023 - May 2025 · 1 yr 8 mos, As a Student Researcher, I worked on the Sign to Speech project — an innovative, funded initiative aimed at enabling two-way communication between sign language users and the hearing population, entirely in Gujarati. Key responsibilities and contributions: • Created a custom Gujarati sign language dataset from scratch, including data collection, labeling, and preprocessing for model training. • Developed a deep learning-based system to convert Gujarati Sign Language into text using custom-trained models. • Implemented speech-to-text translation using AR technology, enhancing the user experience with visual interaction and clarity. • Designed and prototyped a custom hardware device to support real-time translation, enabling a portable, practical solution for daily use. This project deepened my understanding of deep learning, AR integration, and regional language processing, while making meaningful progress toward inclusivity and assistive technology for the Gujarati-speaking deaf and hard-of-hearing community.

Education

Bachelor of Technology - BTech, Computer Engineering, Pandit Deendayal Energy University (PDEU)

Leadership Programme, Business, Management, Marketing, and Related Support Services, Aspire Institute

Public links

github: https://github.com/deepmathukiya

deepmportfolio.web.app: https://deepmportfolio.web.app/