Avinash Maheshwari

Reduced AI response rendering time by 75% via model optimization, fine-tuning, and prompt engineering using TensorFlow/PyTorch. Built LLM-based product recommendation systems for skin & hair analysis. Improved model accuracy by 30% using OpenCV. Deployed 25+ ML models as REST APIs with Vertex AI, AWS SageMaker, and GCP. Led AI teams, implemented CI/CD with Docker, and developed an internal React.js labeling tool, reducing annotation time by 40%. Managed cloud storage with GCS & S3.

• Implemented computer vision models for smart office systems.• Collaborate with cross-functional teams to identify business needs and develop solutions.• Develop algorithms using Natural Language processing and Deep Learning models for recommendation system.• Optimize and fine-tune models for improved performance and accuracy.• Use Git as a version control and development environment (CI / CD in Gitlab, Linux VM).

• Designed and implemented new features for the Corporate Match web application to improve functionality and user experience. • Utilized Rails gems including Capybara, Searchkick, and Filterrific for automated testing and enhancing app performance. • Maintained, troubleshot, and optimized existing software systems to ensure consistent reliability and performance. • Conducted market trend research to inform app development strategies and identify opportunities for feature enhancements and innovation.

Thesis: Consideration of Different Variants of Large Margin Learning Vector Quantization (LMLVQ).