🌟 Electrical Engineering Undergraduate | Passionate about AI, ML, LLMs, Agents, & Intelligent Systems
I’m Hushan Ridmika, an Electrical Engineering undergrad at the University of Moratuwa, ignited by an unquenchable passion for Artificial Intelligence, Machine Learning, Large Language Models (LLMs), Agents, and Intelligent Systems. I’m a visionary builder, crafting AI solutions that fuse groundbreaking innovation with tangible impact. My journey thrives on creativity, technical mastery, and a relentless drive to push boundaries.
- 🤖 LLMs
- 🕸️ Agentic Architectures (CrewAI, LangGraph)
- 📄 RAG, memory chaining, vector DBs
- 🧪 PEFT, LoRA, QLoRA finetuning
- 💡 Real-world, human-centered AI applications.
A full-stack AI solution built for the real world, intelligent, and personalized.
- 👟 Led a 3-member engineering team to build an AI-powered footwear recommendation engine on a Raspberry Pi, using a custom ML pipeline. Integrated OpenCV for real-time outfit color analysis, weather APIs, and occasion-aware logic to deliver fully offline, context-sensitive suggestions.
- 🌧️ Engineered a hybrid wetness detection system, fusing input from a capacitive soil moisture sensor and computer vision-based estimation. This innovation ensured high-accuracy detection of both internal and sole moisture—enhancing automation reliability.
- ⚙️ Designed and implemented a LabVIEW-integrated PID control loop to dynamically adjust a Rigol Ultra Sigma power supply. Orchestrated real-time fan speed modulation, PWM-to-voltage conversion, and precision NEMA motor control—demonstrating seamless AI + control systems integration.
Collaborative intelligence in action—multi-agent orchestration that redefines support.
- 🧩 Designed a modular multi-agent ecosystem using CrewAI, purpose-built to simulate professional-grade support workflows.
- 🗣️ Developed a Senior Support Representative agent integrated with ScrapWebsiteTool to deliver rich, human-like responses with wit, memory, and accuracy.
- ✅ Built a Support Quality Assurance Specialist agent that enforced dialogue correctness using embedded memory, rule-based evaluation, and context tracking.
- 🔁 Together, these agents demonstrated dynamic coordination, memory persistence, and AI-to-AI cooperation—setting a new standard in LLM-powered agentic systems.
- Proven leadership in AI, ML, and IoT innovations.
- Unwavering passion for advancing intelligent systems.
- A collaborative mindset ready to conquer global challenges.
Join me on this AI revolution—let’s build the extraordinary! 🚀