D. James Fusilier

Experience

• Machine Learning Specialist — Invisible Technologies

  Dec 2025 – Present

  New York, NY (Remote)

  • Lead structured evaluation of Virtual Agent performance using standardized ML metrics — Task Completion Rate and Conversational Naturalness across large-scale JSON data logs.
  • Design and iterate on complex reasoning tasks to transition LLMs from baseline solutions to optimized, production-ready workflows.
  • Architect evaluation scripts and reference implementations for new benchmark tasks, identifying SOTA datasets and analyzing frontier research papers.
  • Perform advanced LLM debugging — probing for failures in optimization, fairness, regularization, and transformer architecture; produce detailed failure-mode documentation for applied NLP systems.
  • Deliver structured SFT feedback and refined training data, measurably improving accuracy, reliability, and utility of AI model outputs at scale.

• AI Trainer (RLHF) — Handshake AI Fellow — Handshake AI · University of Houston-Downtown

  Nov 2025 – Present

  San Francisco, CA (Remote)

  • Design and execute end-to-end RLHF pipelines, applying software engineering depth to optimize LLM reasoning, debugging capabilities, and architectural decision-making across Python, JavaScript, and SQL.
  • Develop high-quality Supervised Fine-Tuning (SFT) datasets and preference pairs aligning model outputs with industry-standard software engineering best practices and safety guardrails.
  • Selected as a Software Engineering Expert within the UHD Fellowship program, specializing in RLHF methodology and model alignment.
  • Synthesize code-quality evaluations across multiple programming paradigms, driving measurable improvements in LLM code generation, refactoring, and debugging.

• Senior GenAI / AI Developer & Prompt Engineer — Outlier

  May 2024 – Dec 2025

  Oakland, CA (Remote)

  • Led a team of advanced AI developers designing and deploying scalable GenAI solutions on AWS (SageMaker, Lambda, S3) using Python, TensorFlow, LangChain, FastAPI, and React.
  • Engineered high-performance RAG pipelines with Pandas and vector databases, significantly improving retrieval accuracy and LLM-driven application throughput for enterprise clients.
  • Built responsive full-stack interfaces (React + TypeScript + FastAPI) delivering seamless, high-performance AI-powered enterprise applications.
  • Orchestrated fine-tuned LLM integrations into production systems via FastAPI, boosting operational efficiency and enabling novel AI-driven business workflows.
  • Directed cross-functional Agile teams (Developers, Data Analysts, BI Specialists) to deliver LLM-powered solutions with measurable ROI.

• Senior AI/ML Engineer & Solutions Architect (AI Consultant) — Fishbowl Software

  Oct 2022 – Apr 2024

  Orem, UT (Remote)

  • Designed and deployed a production rule-based chatbot using decision trees integrated with LLM APIs — reduced customer service response times by ~30% for routine inventory queries.
  • Integrated a predictive LLM API into an inventory management platform, enabling data-driven reorder recommendations that significantly minimized stock-out events.
  • Architected the end-to-end ML solution from requirements to deployment, delivering a maintainable, observable system aligned to business KPIs.

• AI Solutions Architect (AI Consultant) — Reserv

  May 2021 – Dec 2021

  New York, NY (Remote)

  • Implemented a predictive maintenance scheduling system via LLM API integration into a property maintenance application — reduced unexpected equipment failures by ~15%.
  • Built CI/CD pipeline for continuous ingestion and analysis of historical maintenance data, enabling real-time equipment failure prediction with automated alerting.

• Mechanical Designer III (Python Automation) — TechnipFMC

  Apr 2013 – Mar 2018

  Houston, TX

  • Integrated NXOpen and TeamCenter using Python and PySpark, streamlining engineering project management workflows and reducing manual overhead for cross-functional teams.
  • Created high-fidelity 3D models and performed finite element analysis (FEA), collaborating across disciplines to deliver precise, compliant engineering solutions.
  • Managed and optimized the enterprise CAD software license portfolio (SaaS/CAE tools), ensuring compliance and maximizing cost efficiency.

Skills

AI / ML & GenAI

RLHF · SFT · Preference Modeling · RAG · Prompt Engineering · LLM Fine-Tuning · Agentic Workflows · Applied NLP · Model Evaluation · MLOps · Benchmark Design

Frameworks & Libraries

TensorFlow · PyTorch · LangGraph · LangChain · Hugging Face · Scikit-learn · NumPy · Pandas · FastAPI · Node.js

LLMs

Claude (Anthropic) · GPT-4o · Gemini · Llama 3.1 · Mistral

Programming Languages

Python · SQL · JavaScript / TypeScript · React · Java · Swift · C# · PHP · HTML / CSS

Cloud & Infrastructure

AWS (SageMaker, EC2, S3, Lambda, CloudWatch) · Azure · Datadog

Databases & Vector Stores

Vector DB · Milvus · Pinecone · MySQL · DB2

Dev Tools

GitHub Copilot · Claude Code · Cursor · GitHub Codespaces · VS Code · Xcode · .NET

Methodologies

Agile / Scrum · CI/CD · Waterfall · Applied AI Engineering

Education

• University of Houston-Downtown — BAAS — Applied Administration · Handshake AI/UHD Fellowship (RLHF & Model Alignment)
• Lone Star College, Houston, TX — AAS — Computer Programming · AAS — Mobile App/Web Development · Computer Programming Specialist II Certification · Computer Programmer Certification