Debbie G.

Key Skills

Software

Other

Top Subject Matter

Clinical audio transcription and classification for patient intake

Real-time sentiment analysis using NLP

AI-assisted document categorization and indexing

Top Data Types

Audio

Text

Top Task Types

Transcription

Emotion Recognition

Classification

Entity (NER) Classification

Question Answering

Freelancer Overview

Software Engineer (Remote) — Built audio transcription and clinical summarization using Whisper/GPT. Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes Master of Science in Computer Science, SUNY, Binghamton University (2024) and Bachelor of Science in Computer Science, University of Alberta (2021). AI-training focus includes data types such as Audio, Text, and Medical and labeling workflows including Transcription, Emotion Recognition, and Classification.

Labeling Experience

Software Engineer - Syndicate-Services LLC

AudioClassificationTranscription

Software Engineer responsible for building automated backend Python pipelines to clean and normalize multi-source vendor data. You handled inconsistent schemas and malformed records using mapping tables and validation rules to improve data quality. You also integrated OpenAI Whisper and GPT to transcribe raw audio and generate structured clinical summaries with dynamic follow-up questions. • Built and normalized vendor datasets with mapping tables and validation logic • Integrated Whisper and GPT for audio transcription and clinical classification • Generated context-aware follow-up questions for streamlined patient intake • Designed workflows for reliable structured outputs from unstructured audio

2025 - 2025

Project — Toxicity analysis with NLP for trend and genre scoring

OtherTextEmotion Recognition

Built a social media trend analysis platform that identifies trending movies/TV shows and measures toxicity across genres. The workflow used sentiment analysis over textual data sourced from Reddit and TMDB APIs. The outcome is labeled toxicity/sentiment metrics enabling comparative genre-level analysis. • Collected social media data via Reddit and TMDB APIs • Performed sentiment analysis to measure toxicity • Produced genre-level toxicity and trend indicators • Built Flask-based APIs for analysis and delivery

2025 - 2025

Project — RAG Q&A pipeline using LangChain and embeddings

OtherTextQuestion Answering

Built a Retrieval-Augmented Generation (RAG) question answering pipeline using LangChain, sentence-transformers, and vector embeddings. The system generated answers grounded in retrieved context, which relies on structured retrieval outputs to support labeled QA pairs and downstream evaluation. This project emphasizes producing model outputs in response to questions using embedded knowledge retrieval. • Implemented RAG using LangChain • Created embeddings with sentence-transformers • Used vector search components (FAISS-style) for retrieval • Generated answers via LLM conditioned on retrieved context

2025 - 2025

Software Engineer (Remote) — Built audio transcription and clinical summarization using Whisper/GPT

OtherAudioTranscription

Created an audio-to-text workflow by integrating OpenAI Whisper to transcribe raw audio and GPT models to classify the content into structured clinical summaries. The pipeline generated context-aware follow-up questions to support patient intake and documentation. This work focused on transforming unstructured audio into labeled, downstream-consumable outputs for clinical use. • Integrated Whisper and GPT for transcription and classification • Produced structured clinical summaries from transcribed audio • Generated context-aware follow-up questions • Built end-to-end pipeline for intake documentation

2025 - 2025

Research Assistant — GenAI categorization and indexing tool for documentation

OtherTextClassification

Developed a smart documentation tool that uses Generative AI to instantly categorize and index user inputs. The system produced classification-oriented labeled outputs to reduce manual data entry. It was deployed as containerized microservices to support real-time search and organizational workflows. • Used Generative AI for instant categorization of user inputs • Generated labeled categories and indexed records • Deployed containerized microservices for real-time search • Reduced manual data entry through automated labeling

2024 - 2025

Show more

Education

Work History