Machine Learning
from Theory to Systems
AI/ML Engineer and Senior Software Engineer working across applied deep learning, computer vision, LLM fine-tuning, retrieval-augmented generation, model evaluation, and cloud-based machine learning workflows.
Selected Work
[ Production-Grade AI Systems ]
Document to Signature Verification Pipeline - Banking Case Study
A professional case study on an end-to-end banking CV pipeline for signature verification. The system converted documents into high-quality image representations, detected signature coordinates with a fine-tuned YOLO model, cropped them without quality loss, cleaned background noise using a fine-tuned U-Net segmentation model, consistent multi-model inference environment, and CNN/Siamese-based similarity models for verification and forgery analysis.
Customer SupportLLM - BaselineSFT with LoRA
Tuned causal LLMs for support via PEFT/LoRA & SFTTrainer. Built data pipelines, BitsAndBytes (4/8-bit) training & batch inference tools for demos.
Security-First RAG Service for Swedish Tax Law
Secure RAG (FastAPI,ChromaDB,SQLite,Fernet), split-storage, redaction. Deploy: Docker, GH Actions, Prometheus, Grafana, Alertmanager, Gradio, HF Spaces.
Expertise
Generative AI & LLMs
LLM fine-tuning, RAG system design, prompt engineering, BERT-family models, Hugging Face ecosystem.
MLOps & Production Serving
FastAPI, Docker, CI/CD, model versioning, evaluation, monitoring, Prometheus/Grafana, Hugging Face Spaces.
Cloud & Data Engineering
AWS SageMaker, Bedrock, Kendra, S3, Kinesis, Glue, Lambda, Data Wrangler, PySpark, data quality workflows.
Computer Vision
Custom CNNs, Siamese networks, YOLO-based detection, U-Net segmentation, OCR, GAN/diffusion synthetic data.
—
Python—
PythonAI Use Case Planner
Python—
Python