Talha Bin Shahid

Talha Bin Shahid

$5/hr
Python Expert, AI/ML Engineer and Data Scientist
Reply rate:
-
Availability:
Full-time (40 hrs/wk)
Location:
Islamabad, Islamabad, Pakistan
Experience:
1 year
TALHA BIN SHAHID AI/ML ENGINEER LinkedIn: raja-talha Github: TalhaBinShahid WORK EXPERIENCE AI/ML Fellowship Jan 2025 - Dec 2025 Buildables | Islamabad Gained hands-on experience in Python (NumPy, Pandas, Matplotlib), web scraping (Selenium, BeautifulSoup, Crawl4AI), and data preprocessing by completing multiple end-to-end projects from scratch. Developed and deployed machine learning and deep learning models (Regression, Classification, Clustering, PCA, CNNs, RNNs, Transformers, GANs) with practical applications such as house price prediction, spam detection, anomaly detection, and image classification, with all projects showcased on GitHub. Designed and implemented advanced AI agents and RAG chatbots using modern frameworks such as LangGraph and the Google AI SDK. PROJECTS LeafLens — End-to-End AI Disease Detection & Assistant (Full-Stack, AI-Integrated): Built an end-to-end plant disease detection system using FastAPI + TensorFlow, fine-tuning ResNet50/InceptionV3 on a 50k+ image dataset and achieving 97–98% accuracy. Developed a full-stack Next.js app with real-time image upload and instant predictions (<2s per image). Integrated an OpenAI/Gemini-powered AI assistant with a RAG-style workflow, improving user decisionmaking speed by 40%+. Pizza Showdown App using Google Agent Development Kit (ADK): Built autonomous agents using Google’s Agentic AI Toolkit and Gemini 2.5 models for web search, translation, and planning tasks Integrated tools like Google Search and structured multi-agent workflows using ReAct and SequentialAgent patterns Managed agent sessions and asynchronous execution with Runner and InMemorySessionService for persistent, context-aware interactions Agentic AI Chatbot using LangChain: Built a locally running AI chatbot using LangChain, integrated with the DeepSeek-R1 model through Ollama, and deployed it via a simple Streamlit interface. Number Plate Recognition Application: Fine tuned YOLO on a dataset of around 25K+ images consisting of Pakistan Images. Extracting Bounding box coordinates of number plates using YOLO. EDUCATION Bachelor of Science in Computer Science National University of Modern Languages, Islamabad, Pakistan CERTIFICATIONS 5-Day AI Agents Intensive Course with Google Machine Learning Specialization by Deeplearning.ai AI for Everyone by Deeplearning .ai Crash Course on Python by Google SKILLS Languages: C/C++, Python, Java, JavaScript, LangChain, LangGraph, crewAI, Google ADK Web Development: FastAPI, React, Next Js Development Tools: Visual Studio Code, Jupyter Notebook, Google Colab, Kaggle Workspace, Google Agent Development Kit. Git & GitHub Machine Learning | Deep Learning | Agentic AI | GenAI | Computer Vision | NLP | Web Scraping Databases: SQL, Vector databases, Faiss, Pinecone, Chroma
Get your freelancer profile up and running. View the step by step guide to set up a freelancer profile so you can land your dream job.