AI Engineering student at ITS · Specializing in Applied ML & LLM Tooling
I learn by building — every project here started as curiosity about how something works. Right now I'm deep into Natural Language Processing, figuring out how machines make sense of human language.
Time-series exploratory analysis evaluating multi-decade historical cycles and geopolitical shifts to identify recurring structural patterns.
Seismic classification pipeline utilizing Gradient Boosting (GBM) on historical earthquake datasets, achieving 93% classification accuracy in predicting tsunami occurrences based on magnitude, focal depth, and epicenter coordinates.
Two-stage hierarchical PyTorch classification pipeline across 5 pulmonary conditions. Isolates normal vs. pathological scans in Stage 1, then mitigates fine-grained cross-entropy confusion between visually adjacent conditions in Stage 2.
Structured LLM agent architecture featuring a 21-protocol cognitive mediation state machine. Implements deterministic communication arbitration, de-escalation heuristics, and strict anti-sycophancy validation without adopting user bias.
A lightweight serverless chatbot that answers questions about this portfolio. Instead of using complex vector databases, it uses a zero-dependency local keyword matching algorithm in Vanilla JS, combined with dynamic live-fetching of GitHub READMEs for project context.
Python-based Model Context Protocol (MCP) server bridging AI assistants with Gym Tracker web app via REST API. Features 31 MCP tools, multi-user sessions, auto-relogin, and deployment on Hugging Face Spaces.
Python-based Model Context Protocol (MCP) server that reverse-engineers myITS Classroom (classroom.its.ac.id) to provide 11 tools for AI assistants to read courses, materials, deadlines, assignments, grades, and download files via MoodleSession authentication.
AI-powered Personal Command Center. Blending a real-time visual life dashboard—spanning from health tracking to academic metrics—with autonomous multi-persona assistants. Engineered for seamless interaction via Web and WhatsApp, complete with daily automation and proactive reminders.
Data-driven experimental platform with continuous prototyping
Data-driven experimental platform with continuous prototyping and deployment.
Some of my proudest recognitions and learning achievements.
Python
PyTorch
Scikit-learn
LLM & Prompt Engineering
Data Analysis
Pandas
NumPy
Git
HTML / CSS
Some of my academic, organizational, and technical experiences.
Listing and finding competitions, helping students find teammates, searching mentors for Students, and handling campus reimbursements.
Focusing on deep learning, computer vision, and scalable AI systems.
Led media strategies and content creation for student organization.
Designed visual identities and managed technical operations for events.
Conducted research and authored scientific papers.
Have a vision? I'm ready to help you bring it to life. I'm always open to discussing new projects, creative ideas, or opportunities to be part of your visions.
Ask me anything about Fahmi's journey, projects, or skills!