•End-to-end Machine Learning Pipeline: Designed and implemented comprehensive data pipelines, incorporating data collection, Roboflow for precision labeling, Python-based preprocessing, and YOLO model training using PyTorch and GCP Object Detection API for scalable deployment.•API Development: Developed and integrated APIs with Flask to deploy machine learning models in production environments, ensuring seamlessmodel integration and real-time inference.•Operational Efficiency Enhancement: Streamlined testing processes, improving operational efficiency by 81% through automated workflows and optimized testing frameworks.•Feature Engineering & Model Optimization: Enhanced existing projects by applying advanced feature engineering and optimization techniques to classical machine learning algorithms (e.g., Random Forest, SVM, XGBoost), reducing bug reports by 45%, and improving model performance, resulting in a 31% increase in user retention.•Cloud Computing & Scalability: Extensive experience in containerizing machine learning models using Docker and deploying them on GKE with GCP, ensuring scalability, reliability, and efficient orchestration for production-level systems.•LLM-Based Generation and Information Retrieval: Utilized LLMs for precise test script generation and information retrieval, leveraging techniques such as prompt engineering and few-shot learning to optimize output relevance.•Model Fine-Tuning for LLMs: Applied advanced fine-tuning techniques, including Parameter-Efficient Fine-Tuning (PEFT), Low-Rank Adaptation (LoRA), and QLoRA, to enhance accuracy and efficiency while minimizing resource usage, ensuring optimal performance for various natural language processing tasks.•Model Quantization for LLMs: Actively involved in quantizing large language models using methods like Post-Training Quantization and Quantization-Aware Training, enabling efficient operation in resource-constrained environments while minimizing performance degradation.

• Spearheaded the end-to-end process of designing, developing, and researching cutting-edge AI systems and models. Conducted comprehensive data collection through meticulous web scraping and sourcing relevant datasets to ensure robust model training. Demonstrated versatility by actively engaging in the continuous improvement of systems through iterative development and retraining, leveraging advanced visualization techniques to derive deeper insights.• Applied a strategic approach to data analysis by meticulously evaluating algorithm use cases, and systematically ranking them based on accuracy and probabilities. Excelled in translating complex algorithmic outputs into insightful visualizations, facilitating enhanced understanding and decision-making. Proactively engaged in system and model optimization, demonstrating a commitment to achieving superior performance in alignment with project objectives.- PPO Offered

• Developed computer vision models to detect plant infestation in real field crop images, utilizingtechniques such as Semantic Segmentation, Edge Detection, and various deep learning architectureswith a focus on analyzing leaf holes as Region of Interest.• Encountered challenges in data quality, biased labels, and limited image preprocessing due tocompression, primarily in "green" color. Hindered by a lack of distinct boundaries amid abundantleaves, no groundbreaking results were achieved. Formulated guidelines for optimal outcomes in thesubsequent data and image collection phase.