AI Engineer | Architecting End-to-End GenAI Systems & Agentic AI
Specializing in Multimodal LLMs, VLMs, and Scalable MLOps/Inference Pipelines

I am an AI Engineer bridging the gap between research and production. I don't just train models; I architect end-to-end AI systems. My expertise spans fine-tuning Multimodal Transformers (VLMs) to deploying scalable inference pipelines using FastAPI, Docker, and Agentic Workflows.

🎓 M.Eng. in AI Engineering for Autonomous Systems
💼 Experience in LLMs, Data Engineering, Predictive Analytics, MLOps, LLMOps, GitHub Actions, Docker, CI/CD
🌱 Always learning and exploring new technologies in AI and ML
📍 Based in Germany | Fluent in English & German (C1)

Deep Learning Architectures:

• 🔷 CNNs (Convolutional Neural Networks) - Image processing & computer vision
• 🔁 RNNs (Recurrent Neural Networks) - Sequential data & time series
• 🔄 LSTMs (Long Short-Term Memory) - Long-term dependencies & memory
• ⚡ Transformers & Attention Mechanisms - State-of-the-art sequence modeling

Advanced AI Models:

• 🤖 LLMs (Large Language Models) - Text generation & understanding
• 👁️ VLMs (Vision-Language Models) - Multimodal understanding
• 🎯 RL Agents (Q-Learning, DQN) - Decision-making systems
• 🔍 Agentic AI - Multi-step reasoning and autonomous agent systems

skills = {
    "AI & ML": ["Deep Learning", "Neural Networks", "LLMs", "Computer Vision", "NLP"],
    "Frameworks": ["PyTorch", "TensorFlow", "Keras", "Transformers", "LangChain", "SmolAgents", "OpenCV"],
    "Data Science": ["Data Analysis", "Statistical Modeling", "Feature Engineering"],
    "MLOps": ["Model Training", "Model Validation", "Model Evaluation", "Model Deployment", "Optimization"],
    "Tools": ["Git", "Docker", "Linux", "Jupyter", "Power BI", "ROS2"]
}

class AIEngineer:
    def __init__(self):
        self.name = "Muhammad Shariq Khan"
        self.role = "AI Engineer"
        self.goal = "Building Systems, Not Just Models"
        
    def daily_workflow(self):
        return [
            "Fine-tune Multimodal Transformers (VLMs)",
            "Design scalable inference pipelines",
            "Deploy production-ready AI systems",
            "Implement Agentic Workflows with Tool Calling"
        ]

Specialized in fine-tuning large language models using parameter-efficient techniques (LoRA/PEFT) for computer vision tasks, achieving significant performance improvements over baseline models.

🎯 Fully functional, tool-augmented AI assistant with autonomous tool-calling (web research, weather, image generation) 🛠️ Tech: SmolAgents, Qwen2.5-Coder, Gemini 2.5 Flash, Streamlit, Docker, GitHub Actions, Chroma DB

🎯 End-to-end predictive system for aircraft turbofan engines using NASA C-MAPSS sensor data 🎯 Predicts Remaining Useful Life (RUL) with 99.15% accuracy using hybrid CNN-LSTM model 🛠️ Tech: Python, TensorFlow/Keras, CNN-LSTM, Pandas

🎯 Embodied AI agent using LLMs as reasoning engine for real-time driver fatigue detection 🎯 Fine-tuned LLaMA 2 for edge deployment with multimodal sensor fusion in CARLA simulator 🛠️ Tech: ROS 2, Meta LLaMA 2, LoRA/PEFT, RAG (FAISS), OpenCV, MediaPipe, CARLA

🎯 Hybrid deep learning framework for vehicle state estimation (position, velocity, orientation) 🎯 Combines RNN, Transformer, and Physics-Informed Neural Networks (PINNs) with Kalman Filter fusion 🛠️ Tech: Python, PyTorch, GRU, Transformers, PINNs

🎯 Deep Reinforcement Learning system optimizing traffic light phases for congestion reduction 🎯 DQN with Experience Replay integrated with SUMO traffic simulator 🛠️ Tech: Python, PyTorch, Deep Q-Network (DQN), SUMO Simulator, NumPy

🎯 End-to-end ML pipeline for large-scale traffic flow analysis using UTD19 dataset 🎯 Custom Neural Network reducing MSE by 61% with K-Means clustering for behavior identification 🛠️ Tech: PyTorch, Scikit-learn, Pandas, K-Means Clustering, Folium

🎯 Self-driving car system using deep learning for steering and speed control from sensor data 🎯 Sensor fusion of Camera + LiDAR with custom CUDA-optimized processing for high-speed inference 🛠️ Tech: ROS 2, PyTorch, Transformers, CUDA, ResNet-18, PointPillars, Open3D

🎯 Comprehensive ML pipeline for house price prediction with automated feature engineering 🎯 Random Forest achieving low error rates with stratified sampling for data reliability 🛠️ Tech: Scikit-Learn, Random Forest, Pandas, NumPy, Joblib

🎯 Lightweight CNN framework built from scratch using C++17 without external ML libraries 🎯 LeNet-5 architecture with INT8 quantization reducing memory by 75%, optimized with Im2col algorithm 🛠️ Tech: C++17, STL, Make, INT8 Quantization, Im2col

🎯 Self-learning autonomous agent for complex grid environments using Standard Q-Learning and DQN 🎯 Built custom "ShariqQuest" environment with comprehensive hyperparameter tuning 🛠️ Tech: Python, PyTorch, Gymnasium, Pygame

🎯 Non-intrusive Driver Monitoring System (DMS) for assessing driver alertness 🎯 Extended 3DGazeNet with eye-behavior analysis and 90.5% gaze classification accuracy 🛠️ Tech: ResNet-18, Geometric Computer Vision, Deep Learning

🎯 Intelligent chat application with persistent message history across sessions. 🎯 Features multi-user support and real-time responses using Gemini 2.5 Flash. 🛠️ Tech: Streamlit, LangChain, MongoDB Atlas, Google Generative AI (Gemini), Python-dotenv

🎯 Fine-tuned Qwen2-VL-7B for thermal camera traffic object counting with 92.76% accuracy. 🎯 Implemented the Sculptor Method for efficient 4-bit LoRA training, reducing VRAM by 43%. 🛠️ Tech: PyTorch, Qwen2-VL, LoRA (PEFT), BitsAndBytes, UrbanIng-V2X Dataset, Docker

✨ AI Research & Development - Working with cutting-edge LLMs and multimodal models
✨ Data Engineering - Building robust pipelines for production ML systems
✨ Predictive Analytics - Condition-based maintenance and reliability optimization
✨ Automation - Creating automated reporting systems and workflows

I'm passionate about staying at the forefront of AI technology:

• 📖 Exploring latest research in LLMs and transformer architectures
• 🧪 Experimenting with new ML frameworks and tools
• 🤝 Contributing to open-source projects
• 💡 Sharing knowledge and learning from the community

I'm always interested in:

• 💬 Discussing AI, ML, and Data Science projects
• 🤝 Collaborating on innovative solutions
• 📚 Sharing knowledge and experiences
• 🌟 Exploring new opportunities

Reach out to me:

• 📧 Email: engr.m.shariqkhan@gmail.com
• 💼 LinkedIn: muhammadshariqkhan
• 🐙 GitHub: @muk0644