Hello, I'm

Syed Affan Hussain

Artificial Intelligence Engineer

 

Passionate AI Engineer specializing in Machine Learning, Computer Vision, Generative AI, and LLMs. Currently pursuing MS in AI at Friedrich-Alexander-Universität Erlangen-Nürnberg, with 4+ years of experience building innovative AI solutions.

Get In Touch
Syed Affan Hussain
4+ Years Experience
20+ Projects Completed
MS AI Student

About Me

Syed Affan Hussain — AI Engineer

Artificial Intelligence Engineer

I am Syed Affan Hussain, a dedicated Artificial Intelligence Engineer with expertise in machine learning, computer vision, and natural language processing. With a Bachelor's degree in Computer Science from Sir Syed University of Engineering and Technology and over 4 years of professional experience in AI development, I bring a comprehensive understanding of modern AI technologies and their practical applications. Currently pursuing a Master's degree in Artificial Intelligence at Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany, I am committed to advancing my expertise while contributing to innovative AI solutions.

Throughout my career, I have specialized in developing sophisticated AI systems for healthcare and security domains, including high-precision disease diagnosis models and intelligent security automation platforms. My technical expertise encompasses end-to-end AI pipeline development, from data preprocessing and model architecture design to deployment and optimization. I have successfully delivered solutions that achieved significant improvements in accuracy, efficiency, and scalability, working with technologies such as deep learning frameworks, computer vision libraries, and large language models.

Beyond technical excellence, I am passionate about knowledge sharing and continuous learning. I actively engage in professional development through certifications from leading institutions including IBM, Stanford University, and Microsoft. My academic achievements include receiving the Valedictorian Award at Sir Syed University and recognition as Best Public Speaker, demonstrating my commitment to both technical mastery and effective communication. I thrive on solving complex real-world challenges through innovative AI applications and collaborative problem-solving approaches.

My Experience

October 2025 - Present

Artificial Intelligence Engineer

EDR Software GmbH

Munich, Bavaria, Germany
  • Built RAG-based chatbots for multiple clients integrating Confluence, Jira, and internal knowledge bases, enabling context-aware responses across distributed documentation systems.
  • Designed document ingestion and embedding pipelines using LangChain and vector databases with chunking strategies to enforce strict multi-tenant data isolation per client.
  • Built an LLM-powered defect classification system that auto-identifies defect categories and extracts relevant metadata from image-based inputs.
  • Reduced hallucinations and improved retrieval accuracy in production by iterating on prompt engineering and chunking strategies across client deployments.
March 2025 - October 2025

Artificial Intelligence Engineer

Lilia AI

Texas, United States
  • Built a multi-agent medical imaging pipeline with LangChain, achieving 92% diagnostic accuracy and cutting radiologist review time by 60%.
  • Developed real-time inference microservices with TensorFlow and FastAPI, delivering explainable diagnostics via Grad-CAM visualizations.
  • Built bias-detection and risk-assessment modules using CPT/ICD-10 coding standards, improving clinical billing efficiency by 40%.
  • Deployed containerized AI systems with JWT authentication in full compliance with HIPAA data protection standards.
March 2023 - March 2025

Artificial Intelligence Engineer

HnH Soft Tech Solutions Pvt (Ltd.)

Karachi, Pakistan
  • Built a LangChain + ChromaDB RAG virtual assistant with persistent memory, serving 100,000+ contextual queries in production.
  • Developed a prostate cancer segmentation pipeline using U-Net++ and ResNet50, achieving a 92% Dice score on MRI data.
  • Built CNN + LSTM anomaly detection models that reduced false positives by 22% in real-world deployments.
  • Set up a CI/CD pipeline with MLflow and GitHub Actions, cutting model deployment cycles from 6 weeks to 7 days.
Nov 2021 - Feb 2023

Artificial Intelligence Engineer

Digital Landscape

Karachi, Pakistan
  • Fine-tuned YOLOv5 on 50K+ annotated images for object detection, reducing false positives by 40%.
  • Built a fraud-detection pipeline using IBM Granite LLMs and transformer-based classification, boosting throughput by 20×.
  • Built a real-time ASL gesture recognition system using MediaPipe, achieving 98% accuracy in live conditions.
  • Deployed Dialogflow CX chatbots to automate patient intake and appointment scheduling workflows.

My Education

April 2025 - Present

Master of Science in Artificial Intelligence

Friedrich-Alexander-Universität Erlangen-Nürnberg

Erlangen, Germany
Jan 2020 - April 2024

Bachelor of Science in Computer Science

Sir Syed University of Engineering and Technology

Karachi, Pakistan

My Skills

Programming

AI / ML Frameworks & Libraries

Agentic AI & LLM Frameworks

Generative AI & LLMs

Web Frameworks & APIs

Cloud & DevOps

Databases

Vector Databases

Development Tools

My Projects

Eventra AI – SOC Assistant

LangGraph LangChain Gemini AI Multi-Agent

Advanced AI-powered Security Operations Center assistant that automates security incident analysis, investigation, and response through a 7-layer multi-agent workflow with 12+ specialized AI agents and human-in-the-loop validation.

12+ AI Agents
7-Layer Architecture
View Project

RxSentinel – Prescription Verification

LangGraph OCR & NLP Gemini AI Healthcare AI

AI-powered prescription verification system leveraging OCR, NLP, and multi-agent workflows to ensure prescription authenticity, DEA/license validation, regulatory compliance, and patient safety with 95%+ accuracy.

95%+ Accuracy
11 AI Agents
View Project

AI Legal Practice Management Platform

FastAPI React 19 + TS AI Agents Multi-Tenant

Full-stack, multi-tenant AI-driven platform for law firms featuring intelligent case analysis, automated document processing, predictive analytics, and semantic search across 4 role-based dashboards (Super Admin, Agency, Legal Agent, Client) with JWT auth and PostgreSQL.

4 User Roles
Multi-Tenant Architecture
View Project

PDF Q&A – RAG Application

LangChain OpenAI RAG Streamlit

Retrieval-Augmented Generation app that lets users upload multiple PDFs and ask natural-language questions. Uses LangChain for document chunking and OpenAI embeddings to deliver accurate, context-grounded answers from the source documents.

Multi-PDF Upload
RAG Pipeline
View Project

LLM Quantization to GGUF

LLM Quantization GGUF llama.cpp

End-to-end guide and implementation for quantizing Large Language Models to the GGUF format using llama.cpp. Shrinks 7B-parameter models from 14GB to ~3.5GB while preserving 95%+ accuracy — enabling fast local inference on consumer GPUs and edge devices.

75% Smaller Size
3-5x Faster Inference
View Project

Brain Tumor Detection System

CNN Python Medical AI Computer Vision

Advanced web-based brain tumor detection system utilizing Convolutional Neural Networks achieving 92% accuracy. Analyzes MRI scans with deep learning techniques to provide reliable medical diagnostic support for healthcare professionals.

92% Accuracy
CNN Algorithm
View Project

Early Disease Detection Platform

Naive Bayes ML Healthcare Classification

Intelligent disease classification system using Multinomial Naive Bayes algorithm. Analyzes patient symptoms from comprehensive dataset covering 150+ medical symptoms for early disease prediction.

150+ Symptoms
ML Approach
View Project

AI-Powered Diet Recommendation

K-Means Clustering Nutrition ML

Personalized nutrition recommendation system using k-means clustering and machine learning. Provides customized diet plans based on user's physical attributes, health goals, and dietary preferences.

AI Powered
Custom Diet Plans
View Project

AI Virtual Gym Trainer

MediaPipe OpenCV Pose Detection Real-time

Real-time AI fitness trainer using pose estimation technology. Provides instant feedback for exercises including Curls, Deadlifts, and Squats with computer vision-based form correction and performance tracking.

3 Exercises
Real-time Feedback
View Project

Particle Swarm Optimization

PSO Feature Selection Optimization ML Enhancement

Advanced feature selection optimization using Particle Swarm Optimization on 17K+ data points. Enhanced Random Forest accuracy from 88% to 93% and SVM from 38% to 98% through intelligent feature engineering.

98% SVM Accuracy
17K+ Data Points
View Project

DALL-E Image Generation Platform

DALL-E OpenAI Streamlit Generative AI

Secure AI-powered image generation platform utilizing OpenAI's DALL-E model. Features user authentication, admin dashboard for request monitoring, and advanced search capabilities for generated content management.

AI Generated
Secure Platform
View Project

Image Segmentation Using Roboflow

Roboflow OpenCV Computer Vision Segmentation

Advanced image segmentation system using Roboflow's pre-trained models for pixel-perfect object boundary detection. Achieves 92.1% pixel accuracy and 85.2% mIoU with applications in medical imaging, autonomous vehicles, and agriculture.

92.1% Pixel Accuracy
85.2% mIoU Score
View Project

My Certifications

Contact Me!