Benjamin Garrard

AI Engineer • Founding Engineer • Systems Builder
Austin, TX
903-830-2348
benjamingarrard5279@gmail.com
GitHub
LinkedIn
bengarrard.com

Summary

AI-focused full-stack engineer and founder with a track record of architecting and shipping complex systems end-to-end. I’ve built document-processing pipelines, vector-search infrastructure, RAG workflows, and intelligent agents using Whisper, Pinecone, Qdrant, OpenWebUI/Ollama, and n8n. As a founder, I’ve owned full product lifecycles—from ingestion and data modeling to real-time interfaces and internal tooling—and delivered production systems where reliability and correctness were non-negotiable. I work best in high-ownership environments where speed, clarity of execution, and deep technical problem-solving matter, and where AI is treated as an engineering multiplier rather than an afterthought.

Core Skills

AI & Knowledge Systems

LLM workflows, prompt chaining, early agent systems
Retrieval-Augmented Generation (RAG), embeddings
Vector stores: Pinecone, Qdrant, OpenWebUI VectorStore
Whisper-based transcription, document & transcript ingestion
Local AI stacks: OpenWebUI, Ollama, n8n AI workflows

Languages & Systems

TypeScript, Python, Go, C#, SQL (Postgres & SQL Server), C++
React/Next.js, Node.js, WebSockets, sockets, async workers
Event-driven design, concurrency, streaming responses, real-time workflows

Architecture & Infrastructure

Cloud Run, Cloud Tasks, microservices, worker queues
SSR/SSG architectures, caching layers, performance tuning
Reporting, auditing, and reliability for revenue & revenue share

Founder-Level Capabilities

Zero-to-one product delivery, ultra-fast iteration, customer-driven development
Operational awareness, field debugging, on-site troubleshooting
Comfort with long hours, shifting priorities, and continuous deployment

Selected Projects

  • Meeting Transcription RAG Pipeline — Used Whisper to transcribe meetings, stored them in a vector database, and built LLM chains to query past discussions and decisions.
  • n8n + Qdrant Ingestion Workflows — Built ingestion pipelines that embed documents and connect LLM nodes to vector search, enabling document-aware automation and chat.
  • Cross-Machine Mouse Sharing (Go + Sockets) — Built a concurrency-driven socket system for real-time mouse transfer between machines.
  • AI Streaming Interfaces — Built streaming-response chat systems for local and remote LLMs to support responsive, iterative workflows.

Education

B.S. Computer Science
Texas State University

GPA 3.7

Additional Work History

Earlier technical roles demonstrating long-term engineering depth, problem-solving, and systems-level thinking — relevant to designing robust AI and tooling.

Python Developer — Charter Communications
2019
  • Migrated APIs from MuleSoft → Python microservices, reducing response time from ~3 minutes to 11 seconds.
  • Built optimized REST patterns and improved reliability for high-volume traffic.
Doctoral Assistant — Texas State University
2018
  • Built PyTorch models achieving ~85% accuracy with fewer than 200 samples.
  • Applied genetic algorithms for performance tuning and hyperparameter search.
Software Developer — Mayachitra, Inc.
2018
  • Authored a Python library (CVJSON) for computer vision data pipelines and dataset management.
  • Built data augmentation + Detectron inference workflows for maritime detection tasks.
VR Lab Software Developer — Texas State University
2017–2018
  • Led a team building VR simulations in Unity3D/C#, including PTSD exposure therapy research tools.
  • Practiced agile development, rapid prototyping, and multi-device testing.

Professional Experience

Founder & CTO — Colorfull
Oct 2023 – Jul 2025
Austin, TX

Built a real-time logistics and ordering platform integrating CloudKitchens' event system, Stripe onboarding, and multi-city operations. Responsible for designing the architecture, maintaining data correctness, and building internal tools for operations and reporting.

  • Designed and operated an async event-driven system (Cloud Run + Cloud Tasks) to process external events reliably and in order.
  • Implemented reporting pipelines to calculate earnings and revenue share for partners and restaurant owners, with automated invoicing.
  • Built internal admin tools for monitoring orders, resolving issues, and managing partner configurations.
  • Used Pinecone for sparse and dense vector workloads around recommendations and search, experimenting with natural language retrieval across internal data.
  • Rebuilt authentication, onboarding, and menu-sync flows days before a critical onboarding event after crashes surfaced, then debugged on-site with restaurant owners.
  • Acted as driver and on-the-ground operator to understand workflows and bottlenecks, then translated those insights into product and backend changes.
Founder — Bashful.ai
Jan 2023 – Oct 2023
Austin, TX

Built a Photoshop plugin in a constrained React/TypeScript environment, integrating AI image generation into creative workflows.

  • Created Python/Flask backend pipelines on GCP for Stable Diffusion jobs, connecting UI actions to AI generation.
  • Designed high-performance UI/animation systems using GSAP + Zustand for responsive creative feedback.
Customer Success Engineer — Moveworks
Sep 2021 – Dec 2022
Enterprise AI Platform

Delivered and optimized AI-powered workflows for enterprise customers; integrated Moveworks' AI platform with their IT systems and data.

  • Shipped frequent updates to customer AI workflows and automations in code, with tight feedback loops from live environments.
  • Integrated with Slack, Teams, Active Directory, ServiceNow, FreshDesk, and NetSuite, connecting AI to real enterprise data.
  • Built an internal Python CLI used by 30+ technical staff that reduced debugging time from hours/days to minutes, saving an estimated $200k.
  • Performed analysis and troubleshooting using ELK, Grafana, and Airflow DAGs, ensuring correctness and reliability of automations.
Software Engineering Consultant — CGI
Jan 2020 – Sep 2021
Consulting / Healthcare Clients

Worked on high-stakes .NET and SQL Server systems for healthcare and enterprise clients where correctness and performance were critical.

  • Normalized large SQL Server datasets and implemented indexing strategies to improve performance and maintain data integrity.
  • Fixed a mission-critical queue system that was incorrectly implemented as LIFO, restoring correct processing order.
  • Architected modernization solutions using .NET, SQL Server, and Azure DevOps for CI/CD.