asad ali

• Responsible for training and refining large language models (LLMs) by designing and curating targeted instruction sets and prompts. Developed tasks specifically to test LLMs capabilities against client-defined success criteria, ensuring model outputs are aligned with desired performance. • Analysed vast datasets to identifying patterns in model failure, using statistical methods to improve success rates. • Business Process Automation: Developed internal Python scripts to automate data entry and reporting workflows, integrating AI models to parse unstructured text into structured databases, reducing manual effort by 40%.

Built and shipped AI-enabled, data-driven web applications using the MERN stack, supporting scalable data ingestion, processing, and low-latency serving. • Designed and delivered backend services and REST APIs powering production NLP chatbot workflows and computer-vision inference pipelines. • Implemented reliable data flows and persistence layers (operational + ML-ready datasets), with focus on performance, scalability, and fault tolerance. • Containerized and deployed services using Docker and CI/ CD to enable repeatable releases and rapid iteration on data/ AI features. • Partnered with product, frontend, and ML teams to convert requirements into deployable, scalable AI solutions and measurable user-facing capabilities

Full-Stack: Architected and delivered custom software solutions for diverse clients, translating abstract requirements into scalable Python applications. Bridged the gap between hardware constraints and software performance using a strong Electrical Engineering foundation. • Algorithm (DSA): Served as a technical mentor for Data Structures and Algorithms (DSA). • Chatbot Development: Deployed custom customer support chatbots using Lang Chain and Hugging Face models, reducing support ticket volume by providing instant AI-driven responses.

• Model Optimization: Engineered and optimized Deep Reinforcement Learning (DRL) agents (SAC/ DDPG) using PyTorch, focusing on hyperparameter tuning and model evaluation. • Data Preparation & Simulation: Managed complex datasets for simulation environments, ensuring high-quality input for model training directly applicable to data preparation pipelines in commercial AI