Robert Gordon
Full-Stack AI Engineer (Python, PyTorch, AWS, TypeScript) | Specializing in MLOps,
Local LLM Deployment & Zero-Trust Infrastructure | USMC Veteran
Summary
Full-Stack AI Engineer with a systems-level approach to building scalable cloud infrastructure and privacy-first AI integrations. Leveraging a decade of USMC operations leadership, I specialize in architecting deterministic, containerized AI microservices using Python, PyTorch, FastAPI, and AWS. Proven ability to bridge the gap between complex enterprise compliance requirements (HIPAA/GDPR) and high-performance MLOps deployments.
Skills
-
AI Systems & MLOps: Python, PyTorch, Hugging Face (PEFT/QLoRA), Local LLM Fine-Tuning, Docker Containerization, NVIDIA CUDA, Vector Databases (Pinecone/Qdrant/PGVector), LLM Evaluation (F1-Scoring, RAGAS)
-
Cloud & Backend: AWS (Lambda, DynamoDB, API Gateway, S3, CloudFront), Node.js, TypeScript, FastAPI, Express.js, RESTful API Design
-
Frontend & Databases: React.js, PostgreSQL, MongoDB, SQL, Tailwind CSS
-
Methodologies: Systems Architecture, CI/CD, Agile-Scrum, Software Development Life Cycle (SDLC)
Work Experience
Independent AI Engineering & Cloud Architecture Immersive
Self-Directed Sabbatical | Nov 2022 – Present- Applied AI Systems & MLOps: Architected and deployed a deterministic, containerized AI microservice for zero-trust PII redaction. Fine-tuned an 8B LLM locally via QLoRA, achieving a perfect 100% Recall rate and 0.86 F1-Score on sensitive enterprise data using FastAPI and direct CUDA GPU passthrough.
- Agentic & Vector Search Pipelines: Built stateful ReAct agent loops and custom HNSW vector search architectures, successfully implementing Semantic ETL, Parent-Child retrieval, and Cross-Encoder re-ranking to eliminate LLM hallucinations.
- Cloud & Serverless Architecture: Earned the AWS Knowledge: Serverless badge. Designed and orchestrated highly scalable, event-driven cloud workflows utilizing AWS Lambda, API Gateway, and Amazon DynamoDB.
- Full-Stack Foundation & Computer Science: Maintained a 3.9 GPA completing advanced B.S. in Computer Science coursework at Park University, strictly focusing on Discrete Mathematics, algorithm efficiency, and full-stack client-server architecture.
Technical Instructor Assistant
Gavlanize Inc. | Aug 2022 – Oct 2022- Supported 60 junior developers by providing guidance on full-stack applications and coding algorithms, resulting in a 50% reduction in project completion time.
- Conducted daily code reviews and weekly study sessions, improving code accuracy by 20% and increasing first-attempt pass rates in the Software Engineer Immersive Residence program by 50%.
- Designed and implemented an onboarding module that reduced new hire ramp-up time by 35%.
Operations Manager
United States Marine Corps | Nov 2018 – Sep 2022- Managed technical, administrative, and training requirements for over 80 personnel, achieving a 100% compliance rating and zero safety incidents.
- Oversaw technical aspects of various programs, leading to a 25% increase in efficiency and a 15% reduction in operational costs through effective problem resolution.
- Developed and implemented the Radar Non-Commissioned Officers Team Leader Course, qualifying over 20 individuals as future Radar Team Leaders and achieving a 100% success rate in their leadership roles.
- Enhanced company readiness by 30% through strategic training initiatives and organizational improvements.
Projects
Architecture of Privacy: Zero-Trust AI Microservice (Capstone)
- Architected a containerized AI microservice using FastAPI and Docker to sanitize unstructured enterprise data (PII), ensuring 100% HIPAA/GDPR compliance by eliminating third-party API data transmission.
- Fine-tuned a local 8B parameter LLM (Mistral/Llama-3) via PyTorch and QLoRA, designing a deterministic regex evaluation pipeline that completely eliminated LLM conversational drift.
- Mathematically benchmarked system safety against a 50-vector golden dataset, achieving a perfect 1.000 Recall (zero data leaks) and 0.86 F1-Score.
Project Ascend: Intracom
- Architected WebSocket real-time communication layer for desktop application enabling live messaging and presence updates across 30+ concurrent users, integrating Socket.IO client-side with React component lifecycle and TypeScript type safety, reducing message latency from polling-based approach by 85%.
- Implemented MongoDB Replica Set infrastructure with Docker containerization and Docker Compose orchestration, establishing production-ready database failover and transaction support for backend API; documented setup procedures enabling team onboarding and reducing deployment configuration time by 60%.
- Built full-stack TypeScript development pipeline spanning Electron/React frontend (18% TypeScript) and Express.js backend (36% TypeScript), implementing automated linting, formatting, and git hooks with Husky and ESLint that improved code consistency and reduced merge conflicts across team contributions.
Cloud-Native Professional Portfolio
- Architected and deployed a high-availability static website using AWS S3 for object storage and AWS CloudFront as a Content Delivery Network (CDN) to ensure sub-second global latency.
- Implemented end-to-end encryption by provisioning a custom SSL/TLS certificate via AWS Certificate Manager (ACM) for the resume.bobbygrdn.com domain.
- Hardened infrastructure security by configuring Origin Access Control (OAC) to enforce the principle of least privilege, restricting S3 bucket access exclusively to CloudFront.
- Managed DNS orchestration via Porkbun, successfully mapping CNAME and ALIAS records for seamless integration between the domain registrar and AWS infrastructure.