I build
AI & ML systems
dependable.
I’m Muhammad Rajif Raditya, an AI & Machine Learning Engineer working at the seam between rigorous research, neural architectures, and useful software.
→ context feedback loop
stays open human-readable boundary
Six projects,
in depth.
Generative AI on edge hardware, RAG retrieval systems, and explainable recommenders — built with TensorFlow, LangChain, Gemini API, and Google Cloud.
More than
the model.
A good prediction is only one component in a system that has to earn trust, carry context, and keep working on an ordinary Tuesday.
Applied Research & Neural Architectures
Translate complex domain challenges into testable neural models, baselines, Neural Collaborative Filtering (NCF) recommenders, and custom embedding pipelines.
Reliable ML Systems & MLOps
Build the connective tissue around a model: data contracts, FastAPI microservices, Docker containerization, Google Cloud Run autoscaling, and automated GitHub Actions CI/CD.
Human-Facing Intelligence & RAG
Design citation-first RAG architectures and feedback loops that expose confidence, provenance, and failure modes instead of asking users to trust a score.
Edge ML & Hardware Integration
Quantize and deploy micro-models (TFLite Micro) on ESP32 microcontrollers for real-time sensor anomaly detection with zero cloud latency.
Integrated Technical Stack & Tools
Machine Learning & Deep Learning
Generative AI, LLM & RAG
Backend, Cloud & MLOps
Where I’ve been working.
UIN Syarif Hidayatullah Jakarta
Information Systems Undergraduate (S.SI)
Pursuing bachelor's degree in Information Systems with an intensive academic focus on Artificial Intelligence, Machine Learning, Neural Network Architectures, Software Engineering, and Database Systems.
Google Skills Arcade Facilitator
Google Cloud
Facilitating hands-on Google Cloud labs and Arcade skill badges for students and developers, guiding learners through cloud infrastructure and GenAI concepts.
Google Student Ambassador
Representing UIN Jakarta to empower the student community by bridging the gap between academics and Google's ecosystem, focusing on AI, Machine Learning, and Google Cloud.
AI/ML Mentor
GDGoC UIN Jakarta
Mentoring students in artificial intelligence and machine learning fundamentals.
AI Engineer Cohort
Pijak in collaboration with IBM SkillsBuild
Intensive training program focusing on AI engineering principles and IBM technologies.
Data Science Mentor
HitLab
Guiding learners through data science projects and methodologies.
A note on how I work.
I am drawn to the point where a model stops being a notebook experiment and starts having to survive in the real world — on a microcontroller with kilobytes of memory, or against a user who asks something the training data never anticipated.
Most of my work sits in that gap: generative AI on edge hardware, retrieval systems that stay honest about their sources, and recommenders whose output can actually be explained. I also teach — mentoring at GDGoC and HitLab has made me a considerably more careful engineer than building alone ever did.
Certifications & Honors
Fifteen completed programmes and 80+ Google Cloud skill badges. The most relevant are listed here.
IBM Granite: Code Generation & Optimization
Issued by: IBM (Sep 2025)
Advanced optimization and code generation techniques using the IBM Granite model. Demonstrated applied knowledge in creating efficient GenAI-driven code tools.
Google Cloud: Develop GenAI Apps with Gemini & Streamlit
Issued by: Google Cloud (Aug 2025)
Built end-to-end Generative AI applications integrating the Gemini API with Streamlit interfaces. Topics included RAG, semantic search, and prompt engineering.
Oracle: Database Programming with SQL
Issued by: Oracle (Dec 2025)
Comprehensive database design and SQL programming, including complex queries, ETL concepts, and highly efficient schema structuring.
JuaraGCP & Google Cloud Arcade Skill Badges
Verified Google Cloud Skills Boost Badge Portfolio
Generative AI & LLM Studio
24 Badges EarnedMachine Learning & Vertex AI
28 Badges EarnedCloud Infrastructure, Docker & Cloud Run
30 Badges EarnedHave a hard
problem?
I like the ones with a fuzzy edge, a real user, and enough room to find a better question.