Dhyaneshvar K

AIML and Data Science Enthusiast

About Me

Hi, I'm a 3rd-year Computer Science student at IIIT Sri City with a deep interest in Artificial Intelligence. I love working with Python and its powerful libraries for data analysis, machine learning, deep learning, computer vision, and natural language processing.

Over time, I've grown especially passionate about Natural Language Processing (NLP), Large Language Models (LLMs), and Reinforcement Learning (RL). Right now, I'm exploring how LLMs can be integrated into real-world applications, while also learning more about web development, model integration, and deployment.

Beyond academics, I'm actively involved in student communities. I was Core Member for AIML & Data Science at Google Developer Groups IIIT Sri City (2024-25) and the current Co-Lead of NexSync (2025-26), where I get to collaborate, build impactful projects, and share knowledge in the AI/ML space.

Projects

Flappy Bird PPO and A2C Agents

A Reinforcement Learning agent that learns to play Flappy Bird using Stable Baselines3. Implements PPO/A2C with entropy annealing for improved exploration.

RL Python CNN PPO A2C OpenCV Numpy

Xview2 Damage Assesment using Siamese UNet

A Siamese UNet with Resnet34 encoder for pre- and post-disaster image segmentation and pixel-wise damage classification. Backend uses FastAPI and Docker-ready.

Python FastAPI PyTorch OpenCV Matplotlib Numpy

HDFS Log Stream Anomaly Detection - Kafka

Built an anomaly detection system for HDFS logs using an Autoencoder. Kafka for real-time log streaming and MLflow for experiment tracking with modular producer-consumer services.

Python TensorFlow Autoencoder Pandas Kafka MLflow Docker Streaming Anomaly Detection

Jarvis - Local personal Multimodal chatbot

A local RAG-based chatbot using Llama-3 models for LLM and vision. Stores and retrieves information via ChromaDB + ElasticSearch with intelligent session management.

LlamaCppPython Python ChromaDB RAG LLM VLM SQLite

GPT2 from scratch using pytorch

Implemented GPT2-small (124M) from tokenizer to attention mechanism. Explored model architecture, token prediction, and training using PyTorch.

LLM Python Pytorch TikToken Transformers

Vision Transformer from scratch for MNIST dataset

Mini Vision Transformer with 4 blocks, 4 attention heads, and 20-dimensional embeddings. Trained on MNIST dataset for image classification.

ViT Python Pytorch Numpy

RAG Enterprise Search

A general RAG chatbot using hybrid retrieval (BM25 + RRF) and cross-encoder reranking. Evaluated on BEIR SciFact dataset

LlamaCppPython Python ChromaDB RAG LLM FastAPI Sentence Transformers Numpy

Bangalore House Price Prediction

My first project ever. Predicts house prices in Bangalore based on area, BHK, bedrooms, and location. Frontend in HTML/CSS/JS, backend in Flask, containerized with Docker.

Scikit-Learn Python HTML CSS JavaScript Flask Matplotlib Docker Pandas

PetVerse - Website for pets

A website for pets with HTML/CSS/JS frontend and NodeJS/Express backend. Uses MongoDB for persistent storage.

HTML CSS JavaScript NodeJS Express MongoDB

AgriIntel - Website for farmers

Helps farmers with crop recommendations, weather forecast, disease detection, yield prediction, and land management using ML models and web interface.

Python Tensorflow Scikit-Learn Streamlit Pandas Prophet Xgboost Matplotlib Seaborn

Skills & Expertise

Technologies and tools I work with to build intelligent systems

Programming & Tools

Python

JavaScript

NodeJS Express HTML/CSS Git Docker Linux

AI & Machine Learning

Deep Learning

Machine Learning

NLP

LLMs

Reinforcement Learning Computer Vision RAG Systems Model Training

Frameworks & Libraries

PyTorch

TensorFlow

Scikit-Learn LangChain FastAPI Streamlit OpenCV Hugging Face Stable Baselines3

Databases & Storage

MongoDB MySQL SQLite ChromaDB ElasticSearch Vector Databases

Data Science & Visualization

Pandas

NumPy

Matplotlib Seaborn Plotly EDA Statistical Analysis

DevOps & Tools

Kafka MLflow Docker Jupyter VS Code Colab Weights & Biases

Certifications

Supervised Machine Learning: Regression and Classification

DeepLearning.AI, Coursera 2024
Unsupervised Learning, Recommenders, Reinforcement Learning

DeepLearning.AI, Coursera 2025

Contact

Interested in collaborating on AI/ML projects? Let's connect!