Designing & Deploying Scalable, Production-Grade ML Systems, Multimodal GenAI, and High-Precision Healthcare Diagnostics.
"Machine Learning models belong in production, not just in Jupyter Notebooks. I bridge the gap between complex statistical research and scalable, fault-tolerant cloud systems."
As an AI/ML Engineer & Healthcare AI Specialist, I focus on building high-throughput ML pipelines, autonomous Agentic AI workflows, and diagnostic computer vision engines. I bring hands-on experience in biomedical data science (Fred Hutch), computer vision deployment (Plus AI), and scalable model development (10Pearls, Neural Seek).
Generative AI • Agentic AI Systems • Retrieval-Augmented Generation (RAG) • Autonomous Agents • LangChain • LlamaIndex • LangGraph • AutoGen • Cohere API • NeuralSeek • Prompt Engineering • Fine-Tuning (LoRA, QLoRA) • Quantization
PyTorch • TensorFlow • Keras • Vision Transformers (ViT) • CNNs • YOLOv8 • Image Segmentation • OpenCV • Hugging Face • LLMs • Multimodal AI • Grad-CAM • NLP
Pinecone • ChromaDB • Qdrant • Milvus • pgvector • Semantic Search • Hybrid Search • Dense Embeddings • Cohere Rerank • BGE Reranker
Supervised & Unsupervised Learning • Feature Engineering • Model Optimization • XGBoost • LightGBM • Reinforcement Learning • Scikit-Learn • NumPy • Pandas • SciPy • Statistical Modeling • Time-Series Forecasting
Python • R • Julia • SQL • SAS Viya • Java • C++ • MongoDB • Jupyter Notebook • VS Code • JetBrains • Matplotlib • Seaborn • Ggplot2
AWS (SageMaker, S3) • Google Cloud Platform (GCP) • Microsoft Azure • Git • GitHub • Docker • FastAPI • Streamlit • MLflow • LangSmith • Weights & Biases (W&B) • CUDA • TensorRT • RESTful APIs
Jira • Trello • Asana • Linear • ClickUp • Notion • Slack • Microsoft Teams • Discord • Miro
- Stack: PyTorch, Vision Transformers, OpenCV, FastAPI, Docker
- Architected deep learning vision models (CNNs & ViT) to classify and segment cancerous lesions from medical scans with zero-loss optimization.
- Built automated image preprocessing & augmentation pipelines in OpenCV, evaluating against clinical metrics (Sensitivity, Specificity, ROC-AUC, F1-Score).
- Stack: Python, LangChain, LangGraph, Pinecone, OpenAI, Streamlit
- Built a multi-agent RAG engine with Pinecone vector database for high-accuracy document indexing and dynamic tool routing via LangGraph.
- Optimized query execution under 800ms response time with real-time UI monitoring in Streamlit.
- Stack: Deep Learning, CNN, OpenCV, Hyperparameter Tuning
- Developed multi-class classification and segmentation pipelines for MRI scans and clinical skin disease images to support early medical diagnosis.
- Stack: Python, Scikit-Learn, Regression Pipelines, Feature Engineering
- Engineered predictive regression models analyzing patient datasets to recommend accurate insulin dosages with minimized RMSE and MAE.
I design, train, and deploy machine learning architectures using a production-first mindset—ensuring scalability, reproducibility, low-latency inference, and continuous monitoring.
- 🥇 Top ML Intern Award – 10Pearls Pakistan
- 🏆 Hackathon Winner – Ghulam Ishaq Khan Institute (GIKI)
- ⭐ Best Student of the Year – Misali School
- 📜 PG Machine Learning Diploma – University of Texas at Austin
- 📜 Maths for Machine Learning Specialization – Imperial College London
- 📜 IBM Certified ML Engineer & SAS Certified ML Engineer
- 📜 Ethical AI (Cambridge) & Applied ML (Alberta ML Institute)
Actively exploring high-impact roles in:
- Machine Learning Engineer / AI Systems Architect
- Data Scientist / Predictive Modeling Specialist
- Healthcare AI / Biomedical Data Researcher
- Computer Vision & GenAI Specialist
- International Remote & On-Site Positions
📩 Direct Email: ramzannaveed659@gmail.com
🔗 LinkedIn: linkedin.com/in/ramzannaveed65
📍 Base Location: Lahore, Pakistan (Open to Relocation)