Jovin Antony Maria

A dedicated and collaborative student with a strong foundation in machine learning and deep learning algorithms. Proficient in Python and TensorFlow, capable of implementing and training various AI models, including computer vision and natural language processing. Demonstrated ability to work effectively as part of a team to achieve project goals.

https://www.linkedin.com/in/jovin-antony-maria-454651326/

Implementation of SSD(Single Shot Multibox Detection) AI algorithm to detect LEGOS

MOST RECENT

• Leveraged TensorFlow and Python to implement and train a Single Shot Multibox Detection (SSD) algorithm.
• Achieved a 93% accurate model used to detect different LEGO pieces from a pile.

AI-Driven Code Generation System

• Developed and implemented a novel AI-powered code generation system using the LlamaIndex framework.
• Coordinated multiple LLMs (Ollama, Mistral, Code_Lama) to generate code from natural language prompts.
• Optimized LLM outputs with RAG for improved code quality and efficiency.
• Created a user-friendly interface for inputting prompts and storing generated code files.

Utilized Ollama and LlamaIndex to create an AI agent for PDF document analysis and response generation.

• Successfully integrated RAG techniques for enhanced information retrieval and generation.
• Demonstrated proficiency in natural language processing and machine learning.

Pipelined MIPS Processor - Computer Architecture Project

• Designed and synthesized a single-cycle MIPS processor using VHDL.
• Successfully implementing it on a Basys3 FPGA board.
• Conducted analysis of the processor's utilization and critical path.

Server-Client Video Streaming

• Developed Python scripts to facilitate video streaming by transmitting MP4 and MP3 files from a server to client computers.