ArielShakaramiro

AI/ML Engineer in progress — LLM, RAG, Computer Vision. Open to full-time AI Engineer roles.

Fine-tuning BERT (bert-base-uncased) for SMS spam/ham classification, benchmarked against a Naive Bayes baseline and BERT as a frozen feature extractor on the same test set. Includes verified evaluation metrics and confusion matrix.

Optimizing MobileNetV2 and YOLOv8n for real-time CPU inference with OpenVINO and INT8 quantization (NNCF), benchmarked on real CCTV footage including a model-limitation case study.

Transformer Encoder for Indonesian intent classification — implementing Positional Encoding & Multi-Head Self-Attention from scratch, compared against fine-tuned BERT via 5-fold cross-validation. Includes overfitting investigation, early stopping, and FastAPI serving.

Character-level vanilla RNN in PyTorch that predicts a name's language of origin, letter by letter. Includes training, confusion matrix evaluation, and inference — plus a real example of dataset distribution shift when tested on modern names.

GPT-2 text generation lab: training a code-completion model from scratch (CodeParrot) and serving it via FastAPI, plus fine-tuning a pretrained Indonesian GPT-2. Results are documented honestly, including undertraining and overfitting findings — not just the wins.