This AI-powered crypto trading bot was built to automate memecoin spot trading on centralized exchanges such as MEXC, leveraging the Hummingbot framework for strategy execution and market connectivity. Owais and his team contributed to designing a modular platform that addressed common challenges in automated trading, including rigid strategy management, lack of real-time visibility, and security risks around exchange API keys. Market analysis highlighted growing demand for always-on trading systems capable of handling high volatility while remaining configurable, observable, and extensible for future AI and DeFi use cases. The solution introduced secure authentication with 2FA, encrypted API key handling, and flexible strategy management through validated YAML files with hot-reload support. Owais and his team helped implement real-time monitoring dashboards displaying PnL, volume, and bot health, alongside live operational logs for rapid debugging and transparency. Each trading bot runs in an isolated Docker container, ensuring scalability and fault isolation, while AI-ready JSON signal injection laid the groundwork for adaptive trading logic. The result was a production-ready MVP that enabled safe, real-time strategy control and positioned the platform for Phase 2 expansion into AI training, futures trading, and decentralized exchange integrations.
Extracted AI – Video-to-Recipe Platform
Extracted AI is a web-based application designed to convert short-form cooking videos from platforms like TikTok, YouTube Shorts, and Instagram Reels into structured, readable recipes. Owais and his team worked on shaping a solution to address a growing usability gap in social-first food content, where visually engaging videos lacked clear ingredient lists, measurements, and step-by-step instructions. Analysis of user behavior showed that viewers frequently struggled to recreate dishes accurately, relying on repeated rewatches, manual note-taking, and guesswork due to the unstructured and fast-paced nature of video content. The platform automates video analysis by extracting captions, metadata, and visual cues, then applying AI models to normalize ingredients, infer quantities, and generate clear cooking steps. Owais and his team emphasized transparency and usability by introducing real-time progress tracking during extraction and parsing, alongside multi-platform support for scalability. The result was a highly efficient workflow that reduced recipe capture time by 90%, delivered complete recipes in under a minute, and transformed entertainment-focused content into practical, reusable cooking instructions—bridging the gap between short-form media and real-world usability.