Harjiven Dodd
AI/LLM Engineer | Systems & Infra | EE + Computational Neuroscience
AI and Machine Learning Engineer with 7+ years of experience specializing in LLM deployment, distributed inference, and AI-driven automation. Built multi-GPU inference clusters for 8–12 clients, production RAG pipelines with RAGAS-evaluated quality metrics, and neural-network memory management systems for continuous-learning AI agents.
Experience & Research
CareerResearch & Projects
Advanced Electrical and Systems Engineer
2025 – PresentLLM-powered document generation, n8n, Microsoft Copilot AI, Azure ML Studio
- Designed AI-augmented document generation workflows using n8n pipelines and Microsoft Copilot AI, significantly reducing manual FMEA and PSA safety analysis creation time
- Built information synthesis systems leveraging LLMs and Azure ML Studio to standardize safety analysis documents (FMEA, PSA, Reliability); integrated with Azure Functions for automated report generation
- Led electrical control systems design, circuit analysis, and comprehensive safety analysis for Honeywell systems
Continuous-Learning AI Agent Memory System
2024 – PresentPersonal Research
- Implemented Titans-MIDAS framework variants for neural-network-based memory management, enabling continuous learning without catastrophic forgetting via surprise-mechanism significance weighting
- Orchestrated multi-step agent workflows using LangGraph with stateful execution graphs and MCP-integrated tool access for external data retrieval and action execution
Titans-MIDASLangGraphMCPPredictive CodingPyTorch
Controls Systems Engineer (Contract)
2025 – PresentAutomated data synthesis — ~$500K annualized savings
- Designed automated data synthesis system consolidating multiple Excel-based mapping documents into comprehensive engineering design documents, saving ~25 hrs/week per engineer (~$500K annualized)
- Developed automation tooling for air-gapped secure foundry environments with no external network dependencies
- Assisted in controls system design and IO mapping for Intel’s upcoming semiconductor fabrication foundry
Distributed LLM Inference Cluster
2023 – PresentPersonal Infrastructure
- Architected multi-machine distributed inference system with Kubernetes orchestration, Docker containerization, load balancing, and FastAPI-served endpoints across GPU-accelerated nodes
- Deployed and optimized 10+ open-source LLMs (LLaMA 3, Mistral/Mixtral, Qwen, DeepSeek, Gemma, Phi) with air-gapped deployment configurations for secure, fully offline inference
KubernetesDockerFastAPIvLLMLLaMA 3Mistral
Advanced Sourcing Engineer
2024 – 2025- Managed vendor/contractor relations and maintained comprehensive supplier databases with performance metrics and cost-benefit analyses across multiple business units
AI Hardware Consulting
2022 – PresentFreelance (8–12 Clients)
- Designed and assembled 8–12 custom multi-GPU and NVIDIA Jetson edge AI systems for LLM inference and training, optimizing configurations across high-throughput (RTX 4090/5090), large-scale training (RTX 6000 Pro, 48GB VRAM), and resource-constrained edge workloads
NVIDIA RTX 4090/5090JetsonMulti-GPUEdge AI
Lead Electrical and Controls Engineer
2021 – 2024Database-driven automated document generation — $150K+ savings
- Led design, integration, and testing of all Safety and Control systems for Plug’s 5MW Peachtree electrolyzer program
- Designed automated document generation system using MS Access, standardizing workflows and eliminating inconsistencies across One Lines, EIDs, Safety Circuit Diagrams, IO Lists, and FATs
- Architected centralized data management system with role-based access controls integrating disparate engineering data sources
Self-Hosted RAG & AI Application Stack
2023 – PresentPersonal Infrastructure
- Built RAG pipelines with RAGAS evaluation (context relevancy, faithfulness, answer relevance) and LLM-as-Judge scoring, served via FastAPI with MCP tool integration and LangGraph agentic retrieval workflows
- Deployed Dockerized AI stack (Perplexica, Paperless-ngx, Nextcloud) with GitHub Actions CI/CD and evaluation regression checks on every commit
RAGASLangGraphFastAPIDockerMCPCI/CD
Electrical Engineer I
2020 – 2021Test station automation — >$200K in savings
- Designed and integrated a Test Engineering station for hardware qualification; identified automation potential, presented initiative to stakeholders, and led development upon approval
- Co-developed automated test execution software and firmware enabling autonomous hardware qualification, delivering >$200,000 in labor and material savings
Deep Learning Video Upscaler
2022 – 2023Neural Networks (GAN, CNN)
- Fine-tuned ESRGAN using LoRA for super-resolution video upscaling on a personally curated dataset, trained on hybrid local multi-GPU and AWS cloud infrastructure (EC2, S3)
ESRGANLoRAAWS EC2S3Multi-GPU
Manufacturer Assembler Specialist
2019 – 2020- Reprogrammed and calibrated manufacturing equipment to improve product quality and reduce production downtime through predictive failure analysis
Advanced Electrical and Systems Engineer
2025 – PresentLLM-powered document generation, n8n, Microsoft Copilot AI, Azure ML Studio
- Designed AI-augmented document generation workflows using n8n pipelines and Microsoft Copilot AI, significantly reducing manual FMEA and PSA safety analysis creation time
- Built information synthesis systems leveraging LLMs and Azure ML Studio to standardize safety analysis documents (FMEA, PSA, Reliability); integrated with Azure Functions for automated report generation
- Led electrical control systems design, circuit analysis, and comprehensive safety analysis for Honeywell systems
Continuous-Learning AI Agent Memory System
2024 – PresentPersonal Research
- Implemented Titans-MIDAS framework variants for neural-network-based memory management, enabling continuous learning without catastrophic forgetting via surprise-mechanism significance weighting
- Orchestrated multi-step agent workflows using LangGraph with stateful execution graphs and MCP-integrated tool access for external data retrieval and action execution
Titans-MIDASLangGraphMCPPredictive CodingPyTorch
Controls Systems Engineer (Contract)
2025 – PresentAutomated data synthesis — ~$500K annualized savings
- Designed automated data synthesis system consolidating multiple Excel-based mapping documents into comprehensive engineering design documents, saving ~25 hrs/week per engineer (~$500K annualized)
- Developed automation tooling for air-gapped secure foundry environments with no external network dependencies
- Assisted in controls system design and IO mapping for Intel’s upcoming semiconductor fabrication foundry
Distributed LLM Inference Cluster
2023 – PresentPersonal Infrastructure
- Architected multi-machine distributed inference system with Kubernetes orchestration, Docker containerization, load balancing, and FastAPI-served endpoints across GPU-accelerated nodes
- Deployed and optimized 10+ open-source LLMs (LLaMA 3, Mistral/Mixtral, Qwen, DeepSeek, Gemma, Phi) with air-gapped deployment configurations for secure, fully offline inference
KubernetesDockerFastAPIvLLMLLaMA 3Mistral
Advanced Sourcing Engineer
2024 – 2025- Managed vendor/contractor relations and maintained comprehensive supplier databases with performance metrics and cost-benefit analyses across multiple business units
AI Hardware Consulting
2022 – PresentFreelance (8–12 Clients)
- Designed and assembled 8–12 custom multi-GPU and NVIDIA Jetson edge AI systems for LLM inference and training, optimizing configurations across high-throughput (RTX 4090/5090), large-scale training (RTX 6000 Pro, 48GB VRAM), and resource-constrained edge workloads
NVIDIA RTX 4090/5090JetsonMulti-GPUEdge AI
Lead Electrical and Controls Engineer
2021 – 2024Database-driven automated document generation — $150K+ savings
- Led design, integration, and testing of all Safety and Control systems for Plug’s 5MW Peachtree electrolyzer program
- Designed automated document generation system using MS Access, standardizing workflows and eliminating inconsistencies across One Lines, EIDs, Safety Circuit Diagrams, IO Lists, and FATs
- Architected centralized data management system with role-based access controls integrating disparate engineering data sources
Self-Hosted RAG & AI Application Stack
2023 – PresentPersonal Infrastructure
- Built RAG pipelines with RAGAS evaluation (context relevancy, faithfulness, answer relevance) and LLM-as-Judge scoring, served via FastAPI with MCP tool integration and LangGraph agentic retrieval workflows
- Deployed Dockerized AI stack (Perplexica, Paperless-ngx, Nextcloud) with GitHub Actions CI/CD and evaluation regression checks on every commit
RAGASLangGraphFastAPIDockerMCPCI/CD
Electrical Engineer I
2020 – 2021Test station automation — >$200K in savings
- Designed and integrated a Test Engineering station for hardware qualification; identified automation potential, presented initiative to stakeholders, and led development upon approval
- Co-developed automated test execution software and firmware enabling autonomous hardware qualification, delivering >$200,000 in labor and material savings
Deep Learning Video Upscaler
2022 – 2023Neural Networks (GAN, CNN)
- Fine-tuned ESRGAN using LoRA for super-resolution video upscaling on a personally curated dataset, trained on hybrid local multi-GPU and AWS cloud infrastructure (EC2, S3)
ESRGANLoRAAWS EC2S3Multi-GPU
Manufacturer Assembler Specialist
2019 – 2020- Reprogrammed and calibrated manufacturing equipment to improve product quality and reduce production downtime through predictive failure analysis
Impact at a Glance
$0K+Total Savings
0+LLMs Deployed
8–12GPU Systems Built
0+Years Experience
Technical Skills
AI/ML & Agents
PyTorchTensorFlowHugging Face TransformersLangChainLangGraphMCP (Model Context Protocol)ESRGANYOLOStable DiffusionWhisperCNNGANDiffusion ModelsRLTitans-MIDASPredictive Coding
LLM Models
LLaMA 2/3Mistral/MixtralQwenDeepSeekGemmaPhiClaude (Anthropic API)GPT-4 (OpenAI API)
Infrastructure & MLOps
vLLMllama.cppDockerKubernetesFastAPIDistributed InferenceMulti-GPU Cluster Managementn8nGitHub Actions (CI/CD)Containerized Model Serving
AI Techniques
Fine-Tuning (LoRA/QLoRA)RAG PipelinesRAGAS EvaluationLLM-as-JudgeVector DatabasesPrompt EngineeringAgentic Workflows
Hardware & Edge
Multi-GPU System DesignDistributed InferenceNVIDIA Jetson Edge AIAir-Gapped AI DeploymentPower Delivery (1–2kW)Thermal Management
Cloud, Platforms & Languages
AWS (S3EC2)Azure (ML StudioFunctions)LinuxGitHubSQLPythonC++C#JavaMATLABVerilogAssembly
Education & Publications
Education
B.S.E. in Electrical Engineering
Arizona State University
Tempe, AZ·May 2019
Publications
Trigeminal Nerve Stimulation in Drug-Resistant Epilepsy: A Systematic Review
Clinical Neurology and Neurosurgery, Vol. 251, April 2025 — April 2025
Co-authors: M.I. Jalal, A.K. Gupta, R. Singh, N.K. Gupta, B. Musmar, A. Singh, D.D. George, M.A. LoPresti, A.M. Wensel
Workforce Trends in Spinal Surgery
World Neurosurgery, September 2021 — September 2021
Co-authors: M.L. Moore, R. Singh, K. McQueen, M.K. Doan, J.L. Makovicka, J.D. Hassebrock, N.P. Patel
Let's build something together
Open to AI/ML engineering roles, consulting, and collaboration.
Secret Clearance — last active August 2021
© 2026 Harjiven Dodd. Built with Astro.