Tariq J.

Overview I build custom AI systems — LLM agents, deep learning models, and NLP solutions — from the ground up. My work spans academic research at Missouri S&T to hands-on freelance delivery for global clients across healthcare, finance, insurance, and edtech. Currently I'm a contractual ML Research Associate at a US University, developing agentic AI for education technology with GradeAnt — AI evaluation of handwritten STEM assignments — and LLM evaluation tasks. Employment ML Research Associate — Missouri S&T, USA | Sep 2024 – Present - Agentic AI R&D for education technology, grant writing, student mentorship Founder / AI Consultant — TensorDot Solutions | Jan 2022 – Present - Custom LLM agents, deep learning, NLP for global clients ML Projects Supervisor (Visiting Faculty) — NED University | Jan 2024 – Present - Supervising 20+ postgraduate AI projects, teaching Deep Learning & GenAI Chapter Lead / MLE — Omdena UAE | Jul 2023 – Jun 2024 - Led global teams on LLM chatbots, medical imaging, vehicle inspection Technical Skills Core: Python, NumPy, Pandas, Scikit-learn Deep Learning: PyTorch, TensorFlow, Transformers, UNet, GANs, LSTMs, CNNs, BLSTM LLMs & Agents: OpenAI SDK, LangChain, LlamaIndex, Ollama, Google AI SDK, ADK, Agno Computer Vision: OpenCV, YOLO, MediaPipe, DeepLab, Vision Transformers Projects - Custom AI Agent Framework — From-scratch agentic AI framework for a client's interactive workshop event. - NLP Curriculum Design — Full NLP bootcamp curriculum for US EdTech startup JadooAI, now open for enrollment. - Voice-Enabled Data Analytics Bot — Falcon/Llama-2 chatbot for voice-driven analytics on Abu Dhabi open data. - ASL Recognition — Transformer-based video classification of 2,000 sign language gestures (PyTorch + TensorFlow). - Legal Document Analysis — Fine-tuned Llama-2 on legal corpus for contract analysis and summarization. - Liver Cancer Detection — HCC detection from histopathology biopsy slides (Omdena). - Cervical Cytology — AI for early disease detection from pap smear images (Omdena). - Vehicle Damage Inspection — CV model for insurance damage detection from photos (Omdena). - Time-Series Forecasting — Hospitality occupancy prediction optimizing pricing and staffing. - BI Analysis — Power BI analysis of UK Fire Brigade station closures for public impact. - Vision Transformer Modeling — Research project using ViT for image classification. - Comparing LLM Agents — Evaluated OpenAI Agent SDK, Google ADK, and Agno for Arabic research work. - Arabic Prompt Optimization — Optimized prompts across multiple LLMs for Arabic NLP tasks. - BLSTM-CNN-CRF Model — Built sequence labeling model for NLP entity extraction. - PyTorch Fine-Tuning — Fine-tuned pre-trained models on small custom datasets (prototype dev). - Custom DL Baseline — Built deep learning baselines from scratch for client comparison studies. - AVA Data Annotation — Managed large-scale Atomic Video Action dataset annotation pipeline. - AI Content Detection — Random Forest model + REST API for AI vs human content classification. - Hate Speech Detection — RNN/LSTM/GloVe based NLP classifier for toxic content moderation. - Audio DeepFake Detection — Deep learning model for synthetic voice and audio forgery detection. - Image Captioning — CNN-LSTM based caption generator on Flickr dataset. - Brain Tumor Detection — Medical image classifier with Streamlit deployment. - Super Resolution — Implemented Swin2SR, EDSR, and LapSRN for image upscaling (8x). - Skeleton Action Recognition — LSTM/GRU/Transformer models for pose-based action recognition. - CCTV Motion Detection — YOLOv3 real-time detection system with Flask backend. - LegalGPT (RAG) — Legal document QA system using LlamaIndex and RAG pipeline. - Medical QA RAG (DR) — RAG system for medical question answering with LlamaIndex + FAISS. - Personal Health Assistant — Medi-Llama based healthcare chatbot for patient interaction. - Ship Detection — Computer vision model for maritime vessel detection. - Audio Transcription (Whisper) — Speech-to-text pipeline using OpenAI Whisper. - Fake Video Detection — Deep learning model for video forgery and deepfake identification. - MRI Classification — Deep learning for medical MRI scan classification. Certifications PG Diploma in Data Science with AI — NED University IBM AI Engineering Professional Certificate Machine Learning (Andrew Ng) — Stanford Online Industrial IoT on GCP — University of Colorado B.E. in Electronics Publications "AI for Aviation" Newsletter — LinkedIn "Application of AI for prediction of asphaltenes stability in crude oils"