Bryanna Elizabeth Binion

AI/ML Software Engineer

PROFESSIONAL EXPERIENCE
Senior AI Full-Stack Engineer Apr 2021 - Present
PreScouter
• Played a key role in back-end development and prompt engineering, leveraging advanced
technologies such as Python, Whisper, and ElevenLabs to optimize system functionality and
performance.
• Developed and fine-tuned a well-trained LLM model using real therapy data within the
psychotherapist chatbot to enhance its conversational capabilities in addressing mental health
concerns effectively.
• Implemented and integrated text-to-speech, speech-to-text, and text-to-text functionalities in real-
time streaming, enabling seamless verbal and text interactions with users for providing therapy and
guidance.
• Enhanced the Chatbot's memory capabilities to store and recall previous conversations, ensuring
personalized and coherent interactions with users during subsequent sessions.
• Leveraged LangChain and LangGraph infrastructures to simplify modifications to the AI model, while
utilizing LLaMa 3.1 hosted on AWS Bedrock to ensure comprehensive user privacy control.
• Migrated to AWS SageMaker for complete control over model training and deployment, while
addressing privacy concerns through a custom LangChain class for streaming outputs, without
altering LangGraph workflows.
• Developed APIs and services using Flask for deploying machine learning models, enabling real-time
predictive capabilities.
• Implemented voice modulation for inputs and outputs, using a faster-whisper solution on AWS
SageMaker for privacy, while leveraging the ElevenLabs API for outputs.
• Created custom memory logic using LangGraph to track conversation flow and reference past
therapy sessions, maintaining an updated client profile.
• Built and improved data pipelines for machine learning workflows using Apache Spark and SQL,
enabling efficient model training and real-time data processing.
• Engineered cloud-based solutions on AWS, utilizing services such as S3, EC2, Lambda, and SageMaker
to deploy scalable and reliable machine learning workflows.
• Explored MLOps strategies by integrating Kubeflow pipelines for end-to-end model lifecycle
management, enabling reproducibility and scalability in deployment.
• Achieved test coverage exceeding 95%, ensuring high quality and reliability in the codebase.
• Integrated PostgreSQL and AWS DynamoDB into the workflow for this functionality.
• Collaborated with the project team to develop a user-friendly front-end interface, facilitating easy
access and navigation for users seeking mental health support.
• Maintained the frontend application with React 17, continuously adding new features using
Bootstrap integration.
• Implemented OpsGenie alerts and DataDog metrics, enabling quick and well-informed decision-
making.
• Technologies Used : Python, Flask, React 17, LLM, PostgreSQL, Docker, Kubernetes, EKS
Machine Learning - Backend Engineer Jan 2016 - Mar 2021
Talent.com | Chicago, IL
• Developed HR Helper Bot on Telegram assists HR professionals by filtering job candidates based on
specified criteria and preferences.
• Processed 2.3M Job Descriptions using fuzzy text matching algorithms and saved $100K automating
manual efforts.
• Built a recommendation engine using Sentence Transformers and Redis, increasing customer
engagement by 40%.
• Performed extensive data preprocessing and quality assurance using Pandas and NumPy, ensuring
data consistency, integrity, and readiness for analysis and model training.
• Implemented secure authentication using OAuth2 and JWT for AI service endpoints.
• Introduced automated ML pipelines with MLflow, streamlining workflows from model
experimentation to deployment.
• Collaborated effectively with product, design, and research teams to align AI functionalities with user
requirements and product goals, fostering synergy and ensuring successful integration of AI
capabilities.
• Leveraged microservice integration and deployment techniques alongside an inference manager,
achieving a 15% improvement in recognition accuracy and a 20% reduction in processing time.
• Maintained and enhanced AI systems in production for reliability, accuracy, and robustness while
utilizing New Relic for comprehensive monitoring to ensure real-time visibility into system
performance.
• Technologies Used : Python, Pandas, Numpy, Redis, MySQL, New Relic
Data Scientist Aug 2019 - Dec 2020
Ecolab | Kennesaw, GA
• Collaborated with teams to build ML model for system stress prediction of projected cost savings of
$1M.
• Utilized VIF scores to detect multicollinearity and systematically eliminating redundant variables.
• Developed guidelines for data preparation for predictive model Crude flex (prediction of oil quality).
• Automated sales performance reports with Advance Excel and VBA, saving 100 monthly manhours.
• Designed PowerBI dashboards with DAX queries to track KPIs, sustaining a 95% customer satisfaction
rate.
• Technologies Used : Python, TensorFlow, StatsModels, Pandas, Scikit-learn, VBA, DAX

Education
Master's Degree in Computer Science, University of Georgia (GPA : 3.84) 06/2017 - 07/2019
Bachelor’s Degree in Computer Science, University of Georgia (GPA : 3.84) 10/2013 - 05/2017
Specialization in Algorithms and AI/ML