Jimmy is available for hire
Hire JimmyJimmy Zebfred
Verified Expert in Engineering
Python Developer
Winchester, TN, United States
Toptal member since July 30, 2025
Bio
Jimmy spent 9+ years building ML/AI solutions for the fintech, healthcare, eCommerce, and software development industries. He has achieved state-of-the-art accuracy while reducing GPU power consumption and operational costs. Jimmy is well-versed in various fields, including data analytics, business intelligence, Agile development, and data engineering.
Portfolio
Toptal
Python, AI Agents, Agentic AI, Amazon Web Services (AWS), Amazon SageMaker...
SambaNova Systems
Artificial Intelligence (AI), Data Science, SQL, Multimodal GenAI, Fine-tuning...
EY
Artificial Intelligence (AI), Data Science, Python, Django, FastAPI, SQL...
Experience
- Python - 9 years
- SQL - 9 years
- Data Science - 8 years
- Machine Learning - 8 years
- Natural Language Processing (NLP) - 6 years
- Azure - 6 years
- AWS Cloud Architecture - 5 years
- Artificial Intelligence (AI) - 3 years
Work Experience
Senior ML/AI Python Developer
2025 - PRESENT
Toptal
- Architected multi-agent harness support (around 12,800 lines of Rust) that extended Maestro's AI SDLC platform from single-agent to running Codex and Gemini alongside Claude Code.
- Delivered Codex and Gemini as production AI agents across 57 tickets — OAuth, credential validation, version pinning, session resume, and Chat-to-CLI sync.
- Optimized the CI/CD pipeline across 38 tickets — Fast CI gates, advisory coverage, per-crate build scoping, E2E sharding — reducing pipeline wall-clock time for 30+ engineers.
- Re-engineered CI cache strategy and artifact pruning, cutting the E2E build cache by around 80% (8.9GB to 1.7GB) and the API cache by around 91% (10.5GB to 0.94GB) to speed cold builds.
- Shipped 194 merged pull requests resolving 152 Jira tickets (around 129,000 net lines) over 8.6 months in a 30+ engineer Rust and TypeScript monorepo.
- Built the shared playbooks-and-skills system, baking reusable agent capabilities into 100% of session images, plus n8n workflow discovery auto-injected at session start.
- Delivered Jira integration end-to-end across 8 tickets: n8n API bridge, Jira CLI access, and full OAuth 2.0 in the UI and session-creation paths.
- Shipped 31 session-lifecycle tickets — fork, purge with orphan detection, container restart on resume, unified stop --all — sustaining 80%+ test coverage.
- Hardened the async build service to classify Cloud Storage and DNS failures, eliminating a class of transient errors that were permanently failing production builds.
- Delivered 22 UI tickets in Maestro's ChatGPT-style redesign — artifacts panel, copy buttons, status indicators — and stood up its Vitest and Playwright suites.
Technologies: Python, AI Agents, Agentic AI, Amazon Web Services (AWS), Amazon SageMaker, Claude Code, AI Pipeline, n8n, Rust, API Connectors, Codex, Google Antigravity, Agent Skills, Claude, Claude plugins, GitHub, CI/CD Pipelines, GitHub Actions, Github Cache, Docker, Kubernetes, Google Cloud Platform (GCP), GSM, Google Kubernetes Engine (GKE), Workflow Automation & System Integration, Prompt Engineering, AI Prompts, TypeScript, React, Git, GitLab, GitHub Runners, Gemini, Amazon Bedrock AgentCore, Agentic RAG Systems, Model Context Protocol (MCP), Langfuse, ElevenLabs Solutions, Speech-to-Text (STT), Agentic AI Systems, Solution Architecture, AI Chatbots, HTML, JavaScript, Troubleshooting
Principal Consultant | Data and AI
2023 - 2025
SambaNova Systems
- Designed and deployed scalable AI solutions for enterprise clients. Managed end-to-end model training, fine-tuning, and deployment, improving inference performance by up to 17% through advanced hyperparameter tuning.
- Integrated advanced model architectures and data processing techniques, working with textual, tabular, and multimodal datasets.
- Led teams and contributed to million-dollar projects while staying aligned with the latest AI research.
Technologies: Artificial Intelligence (AI), Data Science, SQL, Multimodal GenAI, Fine-tuning, BigQuery, Snowflake, Amazon EKS, AWS Glue, AWS Cloud Architecture, Large Language Models (LLMs), Data Engineering, Machine Learning Operations (MLOps), Generative Artificial Intelligence (GenAI), Retrieval-augmented Generation (RAG), LangChain, MySQL, AI Agents, Predictive Modeling, Time Series Forecasting, TensorFlow, Amazon Web Services (AWS), Agentic AI, LangGraph, Prompt Engineering, Pricing, Vector Databases, Data Modeling, AI Pipeline, Architecture, RAG Architecture, Claude Code, Claude, APIs, Workflow Automation & System Integration, AI Automation, API Integration, Software Architecture, PostgreSQL, Model Context Protocol (MCP), HIPAA, Speech-to-Text (STT), Agentic AI Systems, Solution Architecture, AI Chatbots, HTML, Troubleshooting, HubSpot
Senior Data Scientist
2020 - 2023
EY
- Engineered and released LLM-driven applications for Fortune 500 pharmaceutical clients, utilizing RAG and vector database technologies and employing Docker deployments to ensure consistency.
- Integrated UMLS to standardize medical terminologies across clinical datasets, resolving inconsistencies and enabling seamless data integration for improved predictive modeling and streamlined healthcare decision-making.
- Migrated core data processes from Azure Databricks to a Snowflake warehouse, effectively cutting costs down by more than 80%.
Technologies: Artificial Intelligence (AI), Data Science, Python, Django, FastAPI, SQL, Snowflake, Azure Databricks, Amazon SageMaker, Large Language Model Operations (LLMOps), Machine Learning Operations (MLOps), Data Engineering, Data Orchestration, Generative Artificial Intelligence (GenAI), Large Language Models (LLMs), Retrieval-augmented Generation (RAG), LangChain, MySQL, Predictive Modeling, Predictive Analytics, Anomaly Detection, TensorFlow, Amazon Web Services (AWS), Agentic AI, Prompt Engineering, Vector Databases, Neo4j, GraphDB, Data Modeling, Architecture, RAG Architecture, Data Pipelines, APIs, Workflow Automation & System Integration, AI Automation, API Integration, Software Architecture, MongoDB, PostgreSQL, HIPAA, Solution Architecture, HTML, Troubleshooting
Data Scientist, NLP Engineer
2020 - 2020
Ford Motor Company
- Built NLP-based applications using LLMs, Python's NLTK, and SpaCy libraries, enabling sentiment analysis and text classification for improving customer support.
- Analyzed web and mobile usage to determine a prioritized list of suggested intents for chatbot creation, increasing user engagement and serving over 100+ users daily.
- Conducted a comprehensive data analysis on Ford's client data using EDA techniques to uncover insights and inform business strategies, leveraging transfer learning and deep learning for predictive sales modeling.
Technologies: Python, Natural Language Processing (NLP), Machine Learning, Predictive Modeling, Predictive Analytics, Time Series Forecasting, Vector Databases, Data Modeling, Architecture, Data Pipelines, Workflow Automation & System Integration, API Integration, PostgreSQL, HTML, Troubleshooting
Machine Learning Engineer
2019 - 2020
Bank of America
- Introduced an AI-driven system for form filling during account creation, utilizing NLP techniques (Hugging Face's BERT) for text parsing.
- Integrated FuzzyWuzzy for data matching, Google Places API for address autofill, and TensorFlow for model deployment, resulting in a 30% reduction in form completion time and improved user experience and operational efficiency.
- Capitalized on XGBoost tree-based models to classify fraud cases, reducing fraudulent activities by 14%.
Technologies: Data Science, Natural Language Processing (NLP), Anomaly Detection, Python, SQL, Machine Learning, Predictive Modeling, Time Series Forecasting, TensorFlow, Data Modeling, Data Pipelines, APIs, HTML, Troubleshooting
AI Software Engineer
2018 - 2019
Cox
- Applied Hadoop and HBase on Spark using PySpark modules for retrieving data from a NoSQL database.
- Assembled advanced image filters (sharpening, edge detection, denoising) using OpenCV and NumPy, optimizing high-performance image processing, supporting 1,500+ daily active users.
- Integrated AI-driven marketing automation by building predictive models using TensorFlow, enabling targeted customer segmentation and dynamic campaign adjustments.
Technologies: Artificial Intelligence (AI), Data Science, Machine Learning, Computer Vision, NoSQL, Spark, Feature Engineering, Deep Learning, MySQL, Predictive Modeling, Workflow Automation & System Integration, API Integration, HTML, JavaScript, Troubleshooting
Machine Learning Engineer at Lambda School
2016 - 2017
Mozi
- Developed and maintained data pipelines to support data collection, processing, and storage for analysis.
- Labelled and annotated over 1,000 data points to support supervised machine learning models.
- Analyzed large datasets and created clear visualizations to provide actionable insights for business decision-making.
Technologies: Python, Apache Airflow, SQL, Machine Learning, MySQL, Predictive Modeling, Predictive Analytics, Time Series Forecasting, Data Pipelines, APIs, Workflow Automation & System Integration, HTML, JavaScript, Troubleshooting
Project History
MedAI Assist
An AI-powered healthcare automation suite that streamlines clinical documentation, enhances workflow efficiency, and improves decision-making. The system automates clinical documentation by generating SOAP notes and discharge summaries using NLP and speech recognition, reducing physician workload. It also offers medical transcriptions; it converts doctor-patient conversations into structured EHR entries, ensuring accuracy and compliance.
The solution also offers intelligent medical coding: It assigns ICD and CPT codes to records using ML and NLP, improving billing accuracy and reducing errors. It also automates insurance approvals by validating medical necessity, cutting processing time.
The solution also offers intelligent medical coding: It assigns ICD and CPT codes to records using ML and NLP, improving billing accuracy and reducing errors. It also automates insurance approvals by validating medical necessity, cutting processing time.
Skills
Libraries/APIs
TensorFlow, React
Tools
Amazon SageMaker, Codex, GitHub, Apache Airflow, Amazon EKS, Claude Code, Claude, BigQuery, AWS Glue, n8n, Google Kubernetes Engine (GKE), AI Prompts, Git, GitLab
Languages
Python, SQL, HTML, JavaScript, Snowflake, Rust, TypeScript
Frameworks
Django, LangGraph, Spark
Paradigms
Model Context Protocol (MCP), HIPAA Compliance, Anomaly Detection, HL7 FHIR Standard
Platforms
Amazon Web Services (AWS), Docker, Azure, Google Cloud Platform (GCP), Kubernetes, Langfuse, AWS IoT, EPIC Electronic Health Records (EHR), HubSpot
Storage
MySQL, Data Pipelines, PostgreSQL, Neo4j, MongoDB, Redshift, NoSQL
Other
Artificial Intelligence (AI), Data Science, Natural Language Processing (NLP), Machine Learning, Azure Databricks, Large Language Models (LLMs), Retrieval-augmented Generation (RAG), LangChain, Predictive Modeling, Predictive Analytics, Time Series Forecasting, Prompt Engineering, Pricing, Vector Databases, Data Modeling, Architecture, RAG Architecture, Workflow Automation & System Integration, AI Automation, API Integration, Software Architecture, API Connectors, Agent Skills, Claude plugins, CI/CD Pipelines, GitHub Actions, Github Cache, Solution Architecture, Troubleshooting, AWS Cloud Architecture, Healthcare Data Science, FastAPI, Large Language Model Operations (LLMOps), Generative Artificial Intelligence (GenAI), AI Agents, Agentic AI, GraphDB, AI Pipeline, APIs, Amazon Bedrock AgentCore, Gemini, Google Antigravity, Agentic RAG Systems, ElevenLabs Solutions, HIPAA, Speech-to-Text (STT), Agentic AI Systems, AI Chatbots, Fine-tuning, Multimodal GenAI, Data Engineering, Machine Learning Operations (MLOps), Data Orchestration, Computer Vision, Feature Engineering, Deep Learning, GSM, GitHub Runners, RAG Systems
Collaboration That Works
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
Start your risk-free talent trial
Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.
Top talent is in high demand.
Start hiring