🚀 AI/ML Engineer | Data Scientist | Deep Learning Specialist | LLM Expert
AI Engineer driven by results with 9+ months of hands-on experience in Computer Vision, Natural Language Processing, and Deep Learning. Specialized in building and deploying production-ready AI systems using PyTorch, Transformers, and ensemble large language models (LLMs). Proven track record of designing and implementing scalable machine learning pipelines, including reinforcement learning optimization and large-scale computer vision solutions. Strong expertise in end-to-end AI solution development, model deployment, and distributed training across real-world applications.
- 🧠 Build production-ready AI systems with quantifiable business impact
- 🤖 Develop ensemble LLM solutions with reinforcement learning optimization
- 👁️ Create Computer Vision applications processing huge number of frames daily
- 📊 Design scalable ML pipelines with distributed training and deployment
- 🔬 Research early disease diagnosis using advanced computational techniques
- Engineered DragGAN-based image manipulation system, reducing manual editing time by 60% for 500+ daily tasks
- Architected ensemble NLP system integrating 3+ LLMs with RL, improving response accuracy by 35% and reducing hallucination by 28%
- Developed CV solutions achieving 92% accuracy in customer density analysis and 94% precision in parking detection
- Automated 15+ ML workflows using Lightning AI, cutting deployment time from 4 hours to 45 minutes
- Delivered 8+ end-to-end AI solutions (OCR: 95% accuracy, NLP, Data Science) for 5 enterprise clients
- Leveraged GPT-4, Claude, and Llama-2 for production deployments
- Achieved 100% on-time delivery and 25% increase in client retention through direct C-suite collaboration
Programming Languages:
AI/ML Frameworks:
PyTorch • TensorFlow • LightningAI • Scikit-Learn • Hugging Face Transformers • PaddlePaddle
LLM & NLP:
LangChain • LangGraph • CrewAI • LayoutLMv3 • BERT • GPT • Llama • RNN • LSTM
Computer Vision:
YOLO (v5-v8) • OpenCV • MediaPipe • FFmpeg • CNN • Swin Transformer • ResNet
Data Science:
Pandas • NumPy • Matplotlib • Seaborn • Plotly • Jupyter
DevOps & Cloud:
Docker • AWS (EC2, S3, VPC, ECS) • CI/CD • Linux
Web Development:
Django • FastAPI • Flask • REST APIs • Git • GitHub
- Leading 3-member team developing AI model using 5,000+ MRI scans
- Applied KPCA for dimensionality reduction (65% reduction, 98% variance preserved)
- Achieved good classification accuracy** with custom 4-layer CNN (0.723 AUC-ROC)
- Targeting impact on 6.7M+ global Alzheimer's patients
- Built production-ready ensemble integrating OPT-125M, Falcon-7B
- Implemented PPO and Knowledge Transfer Protocol
- Improved response coherence
- Reduced inference latency while maintaining
- 4-bit quantization for flexible GPU/CPU deployment
-
Fine-tuned T5-base on CNN/DailyMail dataset
-
Achieved Good ROUGE-L score
- Developed predictive models using Random Forest and Gradient Boosting
- Achieved R² score of 0.89 on test data
- Deployed with Django REST API
- Implemented using C++ data structures (Quick Sort, Binary Trees, Hashing)
- Optimized search algorithms for O(log n) complexity
Bachelor of Science in Computer Science – FAST-NUCES, Karachi (2021 – 2025)
GPA: 3.5/4.0
Relevant Coursework: Data Structures & Algorithms, Machine Learning, Deep Learning, AI, NLP, Computer Vision, Cloud Computing
- 🏅 Foundations of Data Science – Coursera (2024)
Verify Credential
📧 Email: shahmirmuhammad3@gmail.com
📍 Location: Karachi, Pakistan
📱 Phone: +92 317 2386373
- 🔬 Researched early Alzheimer's diagnosis with 91% accuracy
- 🚀 Reduced ML deployment time from Many hours to almost 2 hours
- 🎯 Processed Surveillance frames daily with CV pipelines
- 📈 Worked on improving chatbot accuracy using ensemble LLMs