Hamza (Ahmed) Khan-Academic Research Contributor in Health Sector (Volunteer)
Check rate
Experience
Academic Research Contributor in Health Sector (Volunteer)
- Acted as technical consultant to optimize multi-layer ensemble models combining ResNet, CNN-BiGRU-Attention, and XGBoost.
- Guided implementation of a Logistic Regression meta-learner to solve class imbalance problems, achieving 92.86% accuracy and 0.9644 AUC on PTB-XL and Chapman-Shaoxing datasets.
Software & AI/ML Engineer
Jurasoft GmbH (RA-Micro)
- Bridged AI backends with vOice interfaces using Node.js and Next.js for real-time AI avatar data flow.
- Developed specialized AI agents: AI Tilda (Translation), AI Peter (Legal), and AI Mary (General Assistant).
- Integrated Whisper AI, WebRTC, and Simli for low-latency, lip-synced AI avatars in virtual conferences.
- Built a privacy-first legal AI platform using local LLMs and custom BERT-based NER models for GDPR-compliant document anonymization.
- Engineered a modular on-device AI deployment architecture using Turbo.net and native macOS packaging.
- Designed and implemented privacy-first legal RAG systems using OpenWebUI, Ollama, FAISS, and Qdrant for semantic retrieval over 8GB+ German legal datasets.
- Developed prompt-based local LLM pipelines using Ollama to extract structured legal information into JSON key-value pairs.
- Implemented contextual retrieval chunking, metadata filtering, semantic search, and class-based indexing strategies to improve retrieval precision and reduce hallucinations.
- Developed custom OpenWebUI tool scripts to dynamically retrieve context from Qdrant because direct custom RAG integration was unavailable in the free OpenWebUI environment.
- Designed prompt engineering and context engineering workflows including legal grounding prompts, structured extraction prompts, and hallucination reduction strategies.
- Created a Scikit-learn document classification system for automated legal email filing workflows.
- Fine-tuned German LLMs (LeoLM and Llama 3.2) using PEFT, LoRA, and quantization techniques.
- Built a specialized C# code converter to automate migration of legacy VB/VBA systems to modern .NET architectures.
- Developed Outlook add-ins using MAUI and Blazor to extend legal software functionality.
Software Engineer
Euronet Worldwide
- Engineered secure backend banking modules using ASP.NET MVC and C# for high-volume financial transactions.
- Developed fraud detection and prevention mechanisms for banking applications.
- Managed complex MS SQL environments using stored procedures and triggers.
- Supported live monitoring and reporting tools for ATM and POS networks with 24/7 availability requirements.
Software Engineer
SalTec Powerlink
- Developed WPF-based industrial configuration software for WIFI, Ethernet, and industrial hardware modules.
- Implemented REST/RPC protocols for remote device management.
- Refactored real-time telemetry systems for Grid, Genset, and UPS power infrastructure.
- Led migration to high-performance database structures using Data Flow Diagrams.
- Collaborated in Agile teams to identify bugs and create test cases for financial applications using JIRA.
Automated Vehicle License Plate Recognition (ALPR)
- Engineered and deployed a YOLOv5-based license plate recognition system optimized for localized Pakistani license plates.
Autonomous Formula F1 Racing Car Detection
- Developed a YOLOv8-based computer vision system for real-time cone, boundary, and trajectory detection for autonomous racing applications.
Industry Experience
See where this freelancer has spent most of their professional time.
Experienced in Information Technology, Professional Services, Healthcare, Energy, and Banking and Finance.
Business Area Experience
See which departments and functions this freelancer has contributed to most.
Experienced in Information Technology, Product Development, Legal, Research and Development, and Quality Assurance.
Summary
AI/ML & Software Engineer specialized in privacy-first intelligent systems, Generative AI, Agentic AI workflows, and German Legal Tech solutions. Experienced in architecting high-performance Retrieval-Augmented Generation (RAG) systems using local LLMs, vector databases, contextual retrieval, prompt engineering, and tool-based AI workflows. Expert in developing custom BERT-based NER models for GDPR-compliant document anonymization and deploying privacy-first AI solutions on-premises. Proven track record leading end-to-end AI integrations—from real-time AI avatars using Whisper AI and WebRTC to local LLM deployment, knowledge retrieval systems, and enterprise software modernization.
Skills
Core Competencies
- Generative Ai (Genai), Agentic Ai, Llm Engineering, Retrieval-Augmented Generation (Rag), Prompt Engineering, Context Engineering, Local Llm Deployment, Nlp & Ner, Computer Vision, Ai System Design, Privacy-First Ai, Backend Engineering, Vector Databases, Ai Evaluation & Llmops
Technical Skills
- Ai Frameworks & Platforms: Chatgpt, Claude, Gemini, Groq, Openai Api, Tavily, Openwebui, Ollama, Langgraph, Langchain, Llamaindex, Huggingface (Peft, Lora, 4-Bit Quantization), Transformers, Whisper-Ai
- Ai & Machine Learning: Gpt-4, Llama 3.2, Leolm (German), Bert, Sentence Transformers, Pytorch, Onnx Runtime, Transfer Learning, Fine-Tuning, Peft, Lora, Quantization
- Llm Engineering & Rag: Rag, Contextual Retrieval, Semantic Search, Prompt Engineering, Context Engineering, Json Knowledge Extraction, Local Llm Deployment, Source Attribution, Hallucination Reduction, Tool Calling, Agentic Ai, Multi-Agent Workflows, Knowledge Retrieval Pipelines, Document Classification, Metadata Filtering
- Ai Evaluation & Llmops: Model Evaluation, Precision, Recall, F1 Score, Confusion Matrix, Retrieval Recall@K, Groundedness Evaluation, Hallucination Detection, Ragas, Deepeval, Langsmith, Mlflow (Working Knowledge), Opentelemetry (Working Knowledge)
- Nlp & Ner: Spacy, Scikit-Learn, Transformers, Flair, Seqeval, Bertfortokenclassification, Named Entity Recognition, Token Classification, Document Classification, German Nlp, Carmen Ner (Per, Loc, Str, Tel, Etc.)
- Vector Databases & Search: Qdrant (Advanced Class-Based Storage), Faiss, Pinecone (Working Knowledge), Vector Embeddings, Semantic Retrieval, Metadata Filtering, Contextual Retrieval
- Supervised Learning: Scikit-Learn, Logistic Regression, Svc, Random Forest, Naive Bayes, Gridsearchcv, Xgboost, Lightgbm, Numpy, Pandas
- Computer Vision: Yolov5, Yolov8, Opencv, Object Detection, Autonomous Vehicle Perception, License Plate Recognition
- Backend & Integration: Python (Fastapi, Flask), Node.Js, C# (.Net Core, Asp.Net Mvc), Java, Rest Apis, Tool Integration, Script Automation, Websockets
- Frontend & Real-Time: Html, Css, Javascript, Typescript, Next.Js, React, Vue.Js (Vuex, Vuetify), Webrtc, Simli (Ai Video), Sambacall, Whisper Ai
- Infrastructure & Tools: Turbo.Net, Docker, Github, Github Actions, Jenkins, Flutter, Ci/Cd, Jira, Linux, Onnx Runtime, Ms Sql, Mysql, Postgresql
- Cloud & Data Engineering Knowledge: Aws (Ec2, S3, Lambda, Bedrock, Sagemaker), Azure (Azure Openai, Azure Ai Foundry, Blob Storage, Virtual Machines), Gcp (Vertex Ai, Compute Engine), Databricks, Apache Spark, Pyspark, Airflow, Dbt, Snowflake, Etl/Elt, Lakehouse Architecture, Kubernetes, Prometheus, Grafana, Arize Phoenix
Languages
Education
Hamdard University
Bachelor of Science, Computer Science · Computer Science · Karachi, Pakistan
Statistics
Experience
Global Experience
Expertise
Qualifications
Profile
Frequently asked questions
Have questions? Find more information here.
Average rates for similar positions
Rates are based on recent contracts and do not include FRATCH margin.
Similar Freelancers
Discover other experts with similar qualifications and experience
Experts recently working on similar projects
Freelancers with hands-on experience in comparable project as a Academic Research Contributor in Health Sector (Volunteer)
Nearby freelancers
Professionals working in or nearby Berlin, Germany
