Open to opportunities

Mohammad Hashemi

@mohammadhashemi

Message

Machine Learning Engineer specializing in computer vision and scalable AI.

Iran

Message

What I'm looking for

I seek roles building production-grade AI/ML systems—especially in CV, LLMs, or forecasting—where I can own end-to-end development, optimize latency, and implement robust monitoring and automation in collaborative teams.

I am a dedicated Machine Learning Engineer with 5+ years of experience in deep learning, computer vision, and backend development, focused on delivering scalable, production-ready AI systems.

I have built real-time forecasting systems, optimized inference pipelines with CUDA and TensorRT, and deployed LLM platforms using LangChain, FAISS, and Milvus to improve factual accuracy and reduce latency.

My projects include YOLOv8-based license plate detection, embedded hand-vein detection on Raspberry Pi, and high-performance C++ optimization algorithms with Python bindings, demonstrating strong end-to-end development and deployment skills.

I prioritize reproducibility, monitoring, and automation—using MLflow, Weights & Biases, Optuna, Docker, CI/CD, and observability stacks—to deliver robust, low-latency solutions for real-world applications.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

TGE

Oct 2025 - Present (8 months)

Developed real-time multi-horizon forecasting systems for BTC, SOL, USDT, and XAU, reducing RMSE ~45% and achieving R² up to 0.97 while optimizing inference latency by ~30% for live trading deployment.

PyTorch CUDA Onnx TensorRT Pandas MLFlow Optuna

AI Engineer

Modsen

Jun 2025 - Oct 2025 (4 months)

Implemented high-performance C++ optimization algorithms for the Maximum Clique Problem and created PyBind11 Python bindings, improving solution quality by 38% and enabling reproducible CI/CD workflows.

Pybind11 C++PyBind11 Genetic Algorithms Tabu Search Simulated Annealing Reinforcement Learning Pytest Docker GitHub Actions

AI Engineer

Rayan Pezhohan

Jan 2025 - Jun 2025 (5 months)

Architected and deployed a scalable real-time LLM platform with RAG (FAISS, Milvus) and hybrid model routing, reducing hallucinations ~45% and cutting end-to-end latency ~30% via caching and optimized pipelines.

LLM Haystack Faiss Milvus fastAPI Docker PostgreSQL Redis Prometheus

Machine Learning & Backend Engineer

Rayan Pezhohan

Nov 2023 - Dec 2024 (1 year 1 month)

Designed and trained YOLOv8-based license plate detection and OCR systems (95% and 92.5% accuracy), optimized training/inference to reduce training time ~60% and supported 100+ concurrent real-time users via Flask backends.

YOLOv8 PyTorch OpenCV Flask PostgreSQL Redis MLFlow CUDA

Computer Vision Engineer

Arooj

Dec 2023 - Nov 2024 (11 months)

Built and optimized hand vein detection and shrimp head cutter systems for embedded devices, achieving U-Net IoU 89% and >95% real-time accuracy while reducing inference time by 30% through quantization and pruning.

U Net PyTorch TensorRT Onnx OpenCV C++Model Quantization