Aaryan Kurade

About Me

I am an ML Engineer specializing in the end-to-end development and deployment of computer vision systems. I have a proven ability to architect robust AI/ML pipelines, containerize models for scalable inference via APIs, and integrate ethical considerations and explainability into the core of a project. I am passionate about building responsible and transparent AI solutions and am actively seeking an internship or full-time ML Engineer role where I can contribute to building transparent and impactful solutions.

Professional Experience

Machine Learning Engineer - Intern | Utopia Optovision Pvt. Ltd.

Jan 2024 - Jan 2025

  • Engineered and Deployed an automated text extraction system using YOLO and PaddleOCR, demonstrating a strong practical understanding of AI/ML concepts and rapid prototyping techniques with Python.
  • Applied advanced computer vision approaches by leveraging OCR and object detection, resulting in a 15% improvement in model accuracy for a key product feature.
  • Streamlined data processing workflows by enhancing technical documentation and automation scripts, reducing manual effort by an estimated 20% and improving team efficiency.

Projects

Deepfake Detection

Engineered a deepfake detection pipeline with PyTorch, achieving 91.2% accuracy. Deployed a scalable FastAPI and Docker-based REST API for real-time inference and integrated Grad-CAM for model explainability, enhancing transparency by visualizing influential facial regions.

PyTorch FastAPI Docker OpenCV XAI
View on GitHub

Video Anomaly Detection System

Developed a production-grade anomaly detection system using a PyTorch-based autoencoder, attaining 92.5% precision on the UCSD Ped2 dataset. Architected a scalable FastAPI and Docker-based REST API, enabling real-time video analysis with a processing speed of ~0.2 seconds per clip.

PyTorch FastAPI Docker MLOps Autoencoder
View on GitHub

AI-Powered Visual Search Engine

Architected a visual search engine for over 100,000 fashion images, delivering results with under 100ms latency. Engineered a full-stack ML pipeline using CLIP and FAISS, optimized with FastAPI and a Streamlit UI, and achieved over 60% memory efficiency with a robust testing framework ensuring 100% code reliability.

CLIP FAISS FastAPI Streamlit PyTorch
View on GitHub

Technical Skills

Education

MIT World Peace University (MIT-WPU) - Pune, India

B.Tech, Electronics and Communication Engineering - AI/ML (Jun 2021 - Jun 2025)