hasanazahid

Backend AI Engineer Intern

500 - Error creating WebGL context.

Jul 2026 - Present • 3 mos

backend engineering , Fast API

GenAI & AI

IT SOLERA

Jun 2026 - Aug 2026 • 2 mos

During my remote Generative AI internship at ITSOLERA (Gen AI Team Beta, June to August 2026), I worked on three sequential projects. First, I distilled diffusion models for real-time BCI EEG denoising, where the CNN student reached 0.40 ms latency, a 2300× speedup, and 94.2% decoding accuracy. Second, I built a RAG-based Opposing-Argument Simulator using the Groq API (Llama 3.1 70B), a FastAPI backend, and a Next.js 14 frontend deployed on Vercel. Third, I fine-tuned Qwen2-VL-2B-Instruct with LoRA and RAG retrieval for hallucination-reduced VQA, tracing root-cause issues to prompt design and the RAG entropy threshold.

Machine Learning Intern

Conneqtor

Feb 2026 - Jul 2026 • 5 mos

Completed 25 machine learning projects covering end-to-end ML workflows. Built and trained deep neural networks for computer vision tasks (YOLOv8, CNNs) with hands-on experience in data preprocessing, model training, and evaluation. Developed supervised learning models (Random Forest, ANN) for regression and forecasting. Implemented unsupervised learning (Isolation Forest anomaly detection, clustering). Worked with reinforcement learning pipelines. Pushed all project code to GitHub repositories. Built interactive Streamlit dashboards and web applications (FastAPI backends, React frontends) to visualize model outputs and metrics. Performed data cleaning, feature engineering, training/validation/test splits, hyperparameter tuning, and performance evaluation across diverse datasets. Gained hands-on experience in the complete ML lifecycle: problem definition → data preprocessing → model architecture design → training → evaluation → deployment. Each project reinforced practical skills in Python, TensorFlow/PyTorch, scikit-learn, and production-ready ML engineering practices.