Shreshth Kapai

Shreshth Kapai

AI Engineer

shreshthkapai@gmail.com +91 6290982150 GitHub

About Me

Btech graduate specializing in Machine Learning, Deep Learning, and AI. Proficient in Python, TensorFlow, PyTorch, generative AI, Machine Learning and Deep Learning. Recognized with nominations for two international conferences, showcasing a proven ability to apply theoretical knowledge to practical projects. Eager to contribute and collaborate in the field of artificial intelligence and machine learning.

Experience

Core Tech Team (Contract)

NeuralThread AI, IIT Bombay Incubator

Aug 2024 - Present | Remote

  • Developed a Minimum Viable Product (MVP) within a month, securing the first client.
  • Fine-tuned, trained, and maintained generative AI models including, but not limited to LLMs, stable diffusion models, and GANs.
  • Designed custom AI workflows.
  • Worked with 3D model generation models like SMPL.
  • Conducted comprehensive comparisons between various GAN architectures.
  • Implemented Docker setups for open-source software like ComfyUI, enabling containerized deployment and scalable execution.
  • Worked with AI interns from IIT Bombay and creative interns from NIFT.

Education

Bachelors of Computer Science Engineering

Vellore Institute of Technology

Sep 2020 - Sep 2024

Skills

Projects

Blog Generation App

Demonstrated LLM proficiency by developing applications for natural language understanding tasks. Leveraged Llama2 and Hugging Face's libraries to create text generation systems, including a blog generation app allowing customization of topic, word count, and audience. Highlighted expertise in advanced NLP techniques and user-centric software design.

Brain Tumor Segmentation

Developed a project utilizing the U-Net architecture for precise delineation of brain tumor boundaries in medical imaging. Improved the accuracy of tumor analysis and diagnosis by leveraging advanced techniques, showcasing expertise in medical image segmentation within this impactful personal initiative.

Automated Medication Scheduling System

Built an AI-powered tool that extracts prescription details using GPT-4o, processes them into structured JSON, and schedules medication reminders, tests, and follow-ups in Google Calendar via its API. Ensured accurate time-based scheduling for improved medication adherence.

RAG Chatbot

Built a chatbot employing Retrieval Augmented Generation (RAG) method. Utilized OpenAI's GPT-3.5-turbo Large Language Model (LLM) via LangChain's ChatOpenAI class, integrated OpenAI's text-embedding-ada-002 for embedding, and leveraged Pinecone vector database as the knowledge base.