Ahamed Hamzad
Computer Science graduate (2026) with experience in Python, machine learning, deep learning, computer vision, NLP, and LLM-powered application development. Built and deployed three AI projects involving model training, RAG pipelines, REST APIs, and cloud deployment on Google Cloud Run.
Research intern at IIIT Allahabad (CVBL) with a project on self-supervised learning under low-label regimes. Certified in Generative AI by NVIDIA DLI.
Experience
Deep Learning Research Intern
Online
Computer Vision & Biometrics Lab (CVBL), IIIT Allahabad Completed DeLTA 2026 — a one-month intensive deep learning program covering CNNs, GANs, RNNs, Transformers, and LLMs organised by IIIT Allahabad. Executed a research project on self-supervised learning: reduced label dependency under low-label regimes of 10%, 20%, and 50% labeled training data; ran experiments across multiple settings and compared results.
Python Development Intern
Shadow Fox
Remote
Built Python automation scripts and REST API integrations; debugged data processing and API response handling pipelines.
Education
BSA Crescent Institute of Science and Technology
Bachelor of Technology
Vandalur, Chennai CGPA: 7.1
Skills
PythonSQLNLPSelf-Supervised LearningRetrieval-Augmented Generation (RAG)YOLOv8FAISSBM25Sentence-TransformersRAGASLLMsTransformersGemini APILang ChainFastAPIREST APIsDockerGoogle Cloud RunGCPGitJupyterStreamlitOpenCVRoboflowVS Code
Projects
Safe Guard
Python, YOLOv8 n, OpenCV, Streamlit, Docker, Google Cloud Run
Fine-tuned YOLOv8 n on a 2,801-image dataset (10 PPE classes); achieved mAP 50: 0.78, precision 0.81; built a real-time detection pipeline with OpenCV and deployed on Google Cloud Run via Docker.
Doc Mind
Python, FAISS, BM 25, Sentence-Transformers, Gemini 2.0 Flash, RAGAS, Streamlit
Built a hybrid RAG pipeline (FAISS + BM 25 + RRF + cross-encoder reranking) exposed as a FastAPI REST API; applied query reformulation to reduce hallucinations; evaluated with RAGAS (faithfulness 0.92, overall 0.85).
Podcast Intelligence
Python, FastAPI, Google Gemini 2.0 Flash, YouTube Transcript API
Built a podcast Q & A system via FastAPI; retrieves and chunks YouTube transcripts using sliding windows; keyword retrieval with term-frequency scoring; Gemini 2.0 Flash generates timestamp-cited, grounded answers.
Certifications
Generative AI & Large Language Models Workshop
IIIT Nagpur & NVIDIA Deep Learning Institute
Machine Intelligence and Vision Algorithms (MIVA-2024)
NIT Kurukshetra