Speech transformer for Japanese emotion recognition
Modified and fine-tuned Wav2Vec2 at the University of Tokyo’s IRCN. First-author paper accepted and presented at ACII 2026.
Jandrik Lana
I develop and evaluate deep-learning models with Python and PyTorch, then build useful applications around them. My work spans speech emotion recognition, computer vision, and an LLM/RAG project.
Quezon City, Philippines
Experience and Education
R&D Contractor · Cocoro EmbodyRemote · Jun 2026–present
Research Intern · University of Tokyo IRCNRemote · Nov 2025–present
AWS Certified AI PractitionerSeptember 2026 
Bachelor of Science in ITTrinity University of Asia
Selected work
Speech and facial emotion estimation
Distilled and evaluated a facial and audio emotion model; developed a Flutter app for parent–infant interaction through multimodal emotion recognition.
Medication identification
Led end-to-end design of a ConvNeXtV2-based mobile vision system with FastAPI on GCP Cloud Run and a Flutter frontend.
Anime recommendation RAG
Built natural-language recommendations over MyAnimeList data using ChromaDB retrieval and Qwen3 served with Ollama.