Barbara V.

Key Skills

Software

Other

Top Subject Matter

Machine learning model training using telemetry and incident reports (livesite incident detection)

Computer vision information extraction from product images

Top Data Types

Text

Image

Top Task Types

Data Collection

Function Calling

Freelancer Overview

Internal ML tool developer (visualization + labeling for ML training) at Microsoft Universal Store. Brings 7+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Other. AI-training focus includes data types such as Text and Image and labeling workflows including Data Collection and Function Calling.

Labeling Experience

Freelance Software Engineer — Groq Llama Vision agent for extracting product attributes

OtherImageFunction Calling

Integrated a Groq Llama Vision AI agent that analyzes product photos in real time. The system automatically extracts product attributes such as names, prices, and color variants from free-form text associated with images. This reduced the need for structured data entry by producing structured outputs from visual and textual inputs. • Processed product photos to derive product-related entities • Extracted structured fields (e.g., name, price, color variants) from unstructured input • Enabled real-time automation for product data generation • Reduced manual uploads and improved publishing workflow speed

2024 - Present

Software Engineer - Microsoft

Image

You worked as a Software Engineer at Microsoft on the Universal Store transaction platform with strict deadlines and high reliability requirements. You developed internal tooling to visualize and label telemetry-derived incident data for machine learning model training and detection. You also delivered Kubernetes-based monitoring and validation tooling to support running transaction services on Azure for many internal teams. • Developed internal tooling to visualize and label telemetry and incident reports for ML model training. • Enabled preemptive detection of livesite incidents before existing alerting systems. • Delivered tooling for monitoring, load testing, and end-to-end validation during service migration. • Supported Azure Kubernetes operation for 50+ teams across Microsoft Commerce.

2020 - 2024

Internal ML tool developer (visualization + labeling for ML training) at Microsoft Universal Store

TextData Collection

Developed an internal machine learning tool that visualized and labeled data using telemetry and incident reports to train machine learning models. The labeled dataset supported preemptive detection of livesite incidents before existing alerting systems. The work involved preparing model-ready labeled inputs derived from operational text-based sources. • Visualized incoming telemetry and incident report content for labeling • Assigned labels to data to create training datasets for ML • Enabled preemptive incident detection via improved model signals • Used tooling workflows to integrate labeled outputs into operational monitoring

2020 - 2020

Education

Work History