AI engineer & lecturer specializing in Computer Vision, NLP, and Time-Series Modeling.
I build end-to-end AI systems from problem definition → implementation → evaluation → deployment, with emphasis on:

• 🧪 Experimental design & model evaluation (Precision, Recall, ROC, HitRate@K)
• 🐳 Reproducible pipelines (Docker-first)
• ⚡ Production constraints (efficiency, maintainability, reliability)
• 📊 Data → Dashboard → Application delivery

• Computer Vision: YOLOv7 / YOLOv8, real-time detection, edge deployment
• NLP / Transformers: BERT / SBERT, semantic similarity, sentiment analysis
• Time-Series: LSTM-based forecasting, multi-source feature integration
• Recommender Systems: hybrid collaborative + content embedding models
• Deployment: Streamlit dashboards, Flask/FastAPI APIs, Dockerized delivery

Tip: setelah README jadi, pin 6 repo yang paling relevan agar terlihat “industrial”.

• 🔍 Hybrid Recommender (NCF + SBERT) — semantic embeddings + collaborative filtering, evaluated with HitRate@K
• 🎥 YOLO Real-time Detection (Edge/TFLite) — optimized inference for deployment environments
• 📈 Sentiment-driven Forecasting — IndoBERT sentiment + LSTM time-series integration
• 🌐 ML Dashboard — Streamlit analytics + Docker packaging
• 🐳 ML Docker Template — reproducible ML repo template (CI-ready)
• ⚙️ API + Model Serving — FastAPI/Flask serving with versioned artifacts

• 🇮🇩 Indonesian — Native
• 🇬🇧 English — Professional
• 🇯🇵 Japanese — JLPT N2 (Near Pass)

Accuracy is not enough.
A model must be reproducible, deployable, and maintainable.

I design AI systems to survive real-world constraints:

• inference speed & resource limits
• monitoring & debugging readiness
• transparent evaluation and iteration

• GitHub: https://github.com/SeedFlora
• Email: budi.juarto@gmail.com