Projects & Work
AI/ML Engineering
Instruction Fine-Tuning
Fine-tuning Large Language Models (LLMs) on specific instruction datasets to improve following capabilities and task alignment.
PyTorch Hugging Face Python NumPyClassification Fine-Tuning
Adapting pre-trained models for specialized text classification tasks with high accuracy on domain-specific data.
PyTorch Hugging Face Python NumPy TensorFlowCoding LLM (End-to-End)
An end-to-end coding LLM built from base, covering tokenization, transformer design, pretraining, fine-tuning, and inference — implemented step by step for clarity and learning.
Building Reasoning LLM Model
Designed and trained a reasoning-oriented LLM architecture using Chain-of-Thought and verifier models.
Auto-Regressive Transformer
Implementation of custom Auto-Regressive Model Transformer Architectures for sequential data processing.
GPT-Style SLM
Trained a GPT-Style Small Language Model (SLM) focused on efficient text generation and low-latency inference.
Ask YouTube
Query YouTube video transcripts in a conversational format using a high-performance RAG pipeline.
GenAI Real World RAG
Developed a real-world Retrieval-Augmented Generation (RAG) system for grounding AI responses in custom data.