Md-SAMI-7

Hi there, I'm Syed Mohammed Samiul Ahmed 👋

🤖 Generative AI & Software Engineer specializing in production-ready RAG applications, high-throughput REST APIs, and microservices.
💼 Former Generative AI Engineer Intern @ BlackBuck Education, engineering GenAI endpoints with FastAPI/Flask and preprocessing 100K+ unstructured data points.
🏗️ Skilled in Cloud & DevOps: Containerizing ML microservices with Docker and deploying on AWS (EC2/S3).
🧠 Solved 500+ DSA problems on LeetCode with strong computer science fundamentals.
🎓 B.Tech CSE @ VIT-AP University (2023 – 2027) | CGPA: 7.87/10.0.
⚡ Motto: Building scalable, high-performance, and impactful AI-driven systems.

Programming Languages

Frameworks & Build Tools

Databases

Cloud & DevOps

Core Concepts & Developer Tools

Feb 2025 – May 2025 | Remote, India

REST API Integration: Designed and integrated high-throughput REST API endpoints using FastAPI and Flask to deploy GenAI and LLM models into production, reducing inference latency by 20% and improving system throughput for cross-functional product integrations.
Python ETL Pipelines: Engineered robust Python ETL pipelines to ingest, clean, and preprocess over 100K+ unstructured data points, improving data-prep efficiency by 35% and informing model training decisions.
Cloud Containerization: Containerized machine learning microservices with Docker, standardizing deployment environments across local development and AWS (EC2/S3) infrastructure, reducing setup time by 40%.

Developed a full-stack RAG web application utilizing LangChain and Streamlit to scrape Wikipedia content, process semantic text chunking, and generate embeddings.
Engineered a hybrid retrieval system using ChromaDB vector store and LangChain's EnsembleRetriever to achieve high-recall context retrieval, improving factual accuracy by ~30% over baseline.
Containerized with Docker for deployment on AWS EC2.

Architected an Applicant Tracking System (ATS) featuring a Python-based resume parsing backend and a React/Redux Toolkit dashboard.
Designed a normalized SQL database schema to store candidate profiles and automated candidate shortlisting with a custom keyword-matching and regex-parsing engine, reducing manual recruiter screening time by 50%.

Developed an Entity Resolution engine in Python using Pandas and NumPy to link and reconcile dirty, unstructured record sets with canonical master datasets.
Implemented character-level n-grams, Jaccard similarity, and Levenshtein distance metrics to resolve company name variations, boosting match rate by 40%.

Designed and trained a Convolutional Neural Network (CNN) on the FER2013 dataset using TensorFlow/Keras to classify 7 distinct facial emotions.
Integrated OpenCV for real-time webcam video feed capture, face detection (Haar Cascades), and frame preprocessing, executing local inference at 30+ FPS.