Luka A.

Key Skills

Software

OpenCV AI Kit (OAK)

Top Subject Matter

Computer vision

Ocr Domain Expertise

and LLM-assisted search for event photo matching.

Top Data Types

Image

Text

Document

Top Task Types

Question Answering

Fine-tuning

Freelancer Overview

Full Stack Developer (AI photo platform) at MiroAI. Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include OpenCV AI Kit (OAK), Internal, and Proprietary Tooling. Education includes Higher Technical Degree in Information Systems, Universidad de Ezeiza (2021). AI-training focus includes data types such as Computer Code and Programming and labeling workflows including Question Answering and Fine-tuning.

Labeling Experience

Full Stack Engineer at BBVA Argentina

Fine-tuning

Integrated a Python machine learning model into a real-time anti-fraud workflow to score customer activity and flag suspicious transactions prior to approval. Built supporting backend services to ingest, clean, and move data between systems for reliable reporting and operational ML-driven features. Added structured logging, request tracing, and observability dashboards to improve reliability of production ML-adjacent pipelines. • Connected an ML scoring model to production banking workflows for real-time transaction risk detection. • Built Python jobs for data cleaning and data movement to support reporting and business decisioning. • Improved system reliability and reduced investigation time via Grafana dashboards and tracing. • Mentored developers on code quality and engineering practices supporting ML integration delivery.

2024 - Present

Full Stack Developer (AI photo platform) at MiroAI

OpenCV AI Kit (OAK)Question Answering

Developed and served AI models for a photo platform where computer vision and OCR outputs drive runner photo matching and data extraction. Implemented Python pipelines and production APIs to support ingestion at scale during traffic spikes. Integrated OCR and computer vision components using OpenCV and PyTorch to enable downstream matching and search features. • Processed and supported computer-vision workloads for image events (e.g., matching photos to runners by bib number and face recognition). • Built FastAPI services exposing vision and OCR model functionality for runner detection. • Packaged machine learning models into reliable production-grade APIs, including model versioning considerations. • Added an LLM-based assistant to enable plain-language searching over photos and event data.

2023 - 2024

Education

Work History