David Thompson
Verified Expert in Engineering
Full-stack Developer
Sudbury, ON, Canada
Toptal member since August 22, 2025
David is a senior Rust engineer with 15 years of experience building production systems. He has designed and shipped tools for complex data workflows and led projects that turn research pipelines into reliable services, workflow engines, parsers for domain languages, and clean APIs. Focused on reliability, performance, and developer experience, David's recent work includes a WDL-inspired executor and agentic AI pipelines that ingested client documents and cut onboarding from 60 days to 1 day.
Portfolio
Experience
- Data Security - 6 years
- Encryption - 6 years
- International Data Privacy Regulations - 5 years
- Async Rust - 5 years
- Rust - 5 years
- Domain-specific Languages (DSLs) - 3 years
- Workflow Description Language - 2 years
- gRPC - 2 years
Preferred Environment
MacOS, Ubuntu, Rust, Python 3
The most amazing...
...thing I've built is a privacy-1st, enterprise-level system that processed video footage to create real-time visitor analytics for commercial real estate firms.
Work Experience
Senior Full-stack engineer, AI/ML Research & Development
Mitratech
- Created agentic AI workflows that reduced customer onboarding from 60 days to 1 day and analyzed 3rd-party vendors for supply chain risks, directly contributing to new and enhanced products.
- Worked on a retrieval-augmented generation (RAG) document analysis pipeline. Engineered a system to auto-fill forms by analyzing complex vendor documents while reducing errors.
- Developed a chatbot, an intelligent conversational interface to improve customer interaction and support.
- Built a vector database-powered RAG system using a hybrid LLM pipeline to analyze SOC 2 deviations from compliance documents, supporting audit processes and automated reporting.
- Designed a contract analysis tool to extract key clauses and dates of enforcement from contracts.
- Trained and deployed models on AWS and Azure to predict the likelihood of a vendor's failure to comply with 3rd-party risk management policies.
- Presented an AI and UX strategy to leadership and design teams to drive adoption of AI-powered platform tools.
- Collaborated with cross-functional teams to translate agentic-AI prototypes into core business products, driving scalability and innovation.
- Prototyped agent workflows using LangChain and custom tooling for evaluating multiple prompts and retrieval strategies across heterogeneous documents.
Senior Rust & Machine Learning Engineer
Freelance Clients
- Designed AI-powered computer vision software for live-streamed events that merges real-time LLM-powered semantic video understanding, multimodal sensor fusion, agentic AI coordination, and decentralized infrastructure.
- Led major developments for real estate companies by transforming retail environments into measurable spaces using real-time computer vision that delivered zone-level foot traffic analytics, dwell time heatmaps, and behavioral funnels.
- Built multi-agent AI workflows for semantic event detection and tracking anonymous human movement through warehouses and mall zones using existing security cameras, all while complying with strict privacy laws.
- Designed and deployed a Rust-based distributed system for real-time video and sensor analytics, scalable, and fault-tolerant 3D digital twin generation.
- Led a global Agile team. Built CI/CD on Kubernetes with Argo CD to scale video stream ingestion.
- Deployed low-latency clients to 5G multi-access edge computing (MEC) nodes using Verizon and AWS Wavelength.
- Built a React front end, Power BI dashboard, and API for real-time analytics and 3rd-party integrations.
- Consulted on applications in real estate, defense, decentralized finance, and smart cities.
Machine Learning Engineer
Deterministic House
- Engineered a generative language model to produce romance novellas and genre fiction using a custom generative adversarial network (GAN).
- Designed a custom vector database to process 3+ million labeled works of fiction.
- Built an internal web app for the editing pipeline and Dockerized digital publishing workflow.
Software Architect
Bitaccess
- Helped create one of the world's first Bitcoin ATM networks. Pioneered methodologies in areas where there were no predefined standards or guidelines.
- Leveraged my deep understanding of blockchain technology, its associated security implications and challenges, and the vision to anticipate its future potential.
- Hired and managed a team of three remote software engineers.
- Developed secure, performance-oriented modules to enable seamless transactions within the Bitcoin network.
Account Manager & Remote Team Lead
Fluidware (acquired by SurveyMonkey)
- Owned a portfolio of enterprise accounts using FluidSurveys and early FluidReview. Translated business goals into clear technical requirements, tickets, and rollout plans.
- Led a small remote team across time zones. Built onboarding checklists, playbooks, and SLAs that improved response quality and reduced handoff friction.
- Implemented light automation for data imports and reporting with SQL and Linux scripting. Reduced repetitive support effort and shortened time to resolution for common requests.
- Partnered with product and engineering to prioritize fixes and features based on quantified customer impact. Drove feedback into roadmap items that improved reliability and usability.
- Created customer-facing guides and admin training material. Improved adoption of advanced features like reviewer workflows and role-based permissions.
Experience
Real-time Computer Vision for Digital Twin Analytics
https://www.grassland.network/The platform ingests video and sensor streams, runs low-latency computer vision and vision-language inference at the 5G edge, and turns scenes into zone-level metrics, dwell heatmaps, and behavior funnels.
Core is a Rust distributed service layer using Tokio, Axum, and SQLx, with per-node caches, backpressure, and graceful recovery. A digital twin model syncs entities in near real time, with tamper-evident provenance via a spatiotemporal blockchain.
I built multi-agent AI workflows in PyTorch and OpenCV, coordinated by task queues and structured logging with tracing. We deployed low-latency clients to Verizon AWS Wavelength and scaled ingestion with Kubernetes, Argo CD, and GitHub Actions.
The product shipped with a React and Power BI front end, OpenAPI-documented APIs, RBAC, and strict audit logging. Outcomes included stable low-latency telemetry, privacy-compliant analytics, and repeatable edge rollouts.
Education
Bachelor's Degree in Computer Science
Queen's University - Kingston, ON, Canada
Skills
Libraries/APIs
OpenCV, OpenAPI, Claude API, NumPy, Node.js, React, REST APIs, API Development, PyTorch, WebGL, TensorFlow, Tokio
Tools
GitHub, Claude, ChatGPT, Claude Code, Amazon EKS, Jira, You Only Look Once (YOLO), n8n
Languages
Rust, Python 3, JavaScript, Domain-specific Languages (DSLs), Python, SQL, HTML, HTML5, PHP, TypeScript, C++, Java
Paradigms
Automation, Model Context Protocol (MCP), Microservices, API Observability, Real-time Systems, Parallel Computing
Platforms
Ubuntu, Linux, AWS IoT, Docker, Kubernetes, Amazon Web Services (AWS), MacOS, Azure, Google Cloud Platform (GCP), AWS Lambda, Amazon EC2, Apache Kafka, Firebase
Storage
PostgreSQL, MySQL, Databases, Redis, Amazon S3 (AWS S3)
Frameworks
Angular, gRPC, Agentic Frameworks
Other
LangChain, Software Engineering, Machine Learning, Artificial Intelligence (AI), Computer Vision, AI Agents, Agentic AI, Stakeholder Management, Async Rust, Encryption, Data Security, API Design, Open Source, APIs, Large Language Models (LLMs), Architecture, Generative Artificial Intelligence (GenAI), Software Architecture, Object Detection, Full-stack Development, Authentication, Back-end, AI Tools, API Integration, Video Processing, Cursor AI, ML Pipelines, Data Analytics, Data Reporting, Image Recognition, Data Engineering, AI Design, AI Programming, Full-stack, Planning, Agentic AI Systems, Agentic RAG Systems, Argo CD, Vector Databases, Retrieval-augmented Generation (RAG), Deep Learning, Workflow Description Language, International Data Privacy Regulations, Supabase, Data Science, Augmented Reality (AR), Machine Learning Operations (MLOps), OpenAI, CI/CD Pipelines, Algorithms, Data Structures, Computer Architecture, Operating Systems, Data Management, Networking, Distributed Systems, Programming Languages, Artificial Neural Networks (ANN), Data Analysis, 5G, Edge Computing, Edge AI, Multiplexers (MUX), Video Streaming, Amazon Kinesis, Amazon Kinesis Video Streams, Geospatial Data, Geospatial Analytics, Privacy, Data Privacy, General Data Protection Regulation (GDPR), YOLOv5, 3D Tracking, Low Latency, Ultra-low Latency Networking, Real-time Streaming, Real-time Data, CMOS Image Sensors, Concurrency, Time Series Data, Time Series Analysis, WebSockets, Load Balancers, Pub/Sub, Know Your Customer (KYC), Scripting, Issue Tracking, Software Development Lifecycle (SDLC), Business Requirements, User Requirements, Async/Await, Axum, Serializer/Deserializer (SerDes), Auditing, Finance, Large Language Model Operations (LLMOps)
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring