Muhammad Ahmed Suhail

Machine Learning Engineer

muhammadahmedsuhail@gmail.comLinkedInGitHubMedium

Demos

Agentic Pair Programming Assistant

Built an agentic network using Google ADK Agents for pair programming with codebase Q&A and automated feature creation, using Qdrant for vector embeddings and FastAPI with Streamlit for the interface.

Google ADKQdrantFastAPIStreamlit

AI-Based Resume Screening System

Built an intelligent candidate shortlisting platform using NextJS, FastAPI, and Supabase that processes high-volume applications through semantic resume analysis with OpenRouter LLM, storing structured summaries as vector embeddings in Qdrant for similarity search.

NextJSFastAPISupabaseQdrantLLM

About

I'm a Machine Learning Engineer with a Bachelor's degree in Data Science from FAST-NUCES. I specialize in building AI-powered systems, from personalized recommendation engines and intelligent chatbots to sentiment analysis pipelines and automated screening platforms.

Based in Islamabad, Pakistan, I enjoy working at the intersection of data science and practical software engineering, turning complex data problems into production-ready solutions.

Skills

PythonMySQLMachine LearningArtificial IntelligenceScikit-learnPyTorchPandasNumPyMatplotlibSeabornPlotlyGitDockerReactFastAPITableauMERNNextJSTypeScriptRESTful APIs

Experience

Data Scientist

Jan 2026 – Apr 2026

Etra Ventures

  • •Built a personalized video ranking microservice for a food-focused short-video platform using Python, FastAPI, and Supabase, deployed as a serverless API on Vercel.
  • •Designed a hybrid recommendation engine with a parent/child architecture that classifies users into lifecycle states (cold start, learning, personalized, re-engagement) based on engagement metrics, running specialized scoring strategies in parallel.
  • •Fused scores via weighted min-max normalization and implemented optional re-ranking using cosine similarity over precomputed user and video embeddings.

Freelance Machine Learning Engineer

Sep 2024 – Jan 2026
  • •Built an AI E-commerce Chatbot with NextJS, FastAPI, and Supabase featuring order tracking, product Q&A, complaint filing, automated human escalation, and a management dashboard, handling 1,000+ monthly interactions and reducing support workload by 80%.
  • •Developed a cryptocurrency sentiment analysis system scraping 10,000+ daily tweets across 15+ coins, correlating with real-time market data using weighted algorithms to identify 10+ credible handles, built with FastAPI and MongoDB.

Projects

Near Real-Time DataWarehouse Analysis

Implemented the MESHJOIN algorithm in Java to build a near real-time ETL pipeline for a shopping store, probing hash tables to stream customer transactions into the Data Warehouse as they arrive. Designed a star schema with facts, dimensions, and aggregated data sources, then created and populated a MySQL database with 10,000+ transactions to optimize selling strategies via data visualization and reporting.

JavaMySQLETLData WarehouseStar Schema

AI-Based Resume Screening System

Built a parallel resume processing app handling 500+ resumes simultaneously, using OpenRouter LLMs for semantic analysis and Qdrant for vector storage with experience-weighted filtering per job role. Implemented an agentic retrieval system that performs semantic search over candidate embeddings, ranking shortlists based on role-specific experience requirements for fully automated screening.

PythonOpenRouterQdrantLLMFastAPI

Education

Bachelor of Data Science

2020 – 2024

FAST-NUCES

CGPA: 3.37

Relevant Coursework: Probability & Statistics, Big Data Analytics, Data Visualization, Data Warehousing & BI, Artificial Intelligence, Data Mining, Deep Learning, MLOps

Honors & Awards

  • ◆Third position in academic standings for the seventh semester
  • ◆Dean's List for the sixth, seventh, and eighth semesters