- 🌱 Currently learning with freeCodeCamp's certification paths and TechLabs Berlin' Deep Learning track
- 🔭 Working on SAMO as the sole Deep Learning Engineer - An AI-powered, voice-first journaling companion providing real emotional reflection through advanced emotion detection and natural language understanding:
- Independently managed complete ML lifecycle from dataset analysis and model experimentation in Google Colab to production deployment on cloud infrastructure.
- Unified pipeline processing voice, text, and emotional data with seamless model chaining for complex NLP tasks.
- Production-grade emotion detection using fine-tuned DistilRoBERTa (90.70% accuracy) + T5 summarization + OpenAI Whisper voice processing in unified pipeline achieving <500ms end-to-end latency.
- Deployed scalable MLOps infrastructure on GCP with Docker microservices, Prometheus monitoring, automated testing, and ONNX optimization (2.3x speedup, 75% model compression).
- Stack: PyTorch, Transformers, ONNX Runtime, Docker, Kubernetes, GCP, Flask APIs with expertise in transformer architectures, model optimization, and production ML deployment.
