Selected Work

Production systems and applied AI projects built around real operational problems.

EXPLORE PROJECTS

CUSTOM PLATFORMS

Production systems built around the real workflows of the businesses that use them.

Client System · In Production

Field Operations and Payroll for a 50-Person Delivery Business

A custom platform on top of the dispatch software the company already uses, covering manifests, crew check-in and check-out, hours, tips, payroll and invoicing. We built it, we own the infrastructure it runs on, and we ship changes from the owners' ticket queue.

SEE THE PLATFORM →
Field operations platform — per-person tools
In Production

Listing Media Center — An Operations Platform for a Real Estate Media Company

A custom platform that runs bookings, scheduling across a fleet of photographers, delivery, invoicing and contractor payouts in one place. Built over several years, in production today, and used by multiple brokerages.

SEE THE PLATFORM →
Listing Media Center — projects dashboard

APPLIED AI IN OPERATIONS

Practical examples of how we apply AI to everyday operational work—reading documents, interpreting images and video, finding information, and helping people work more efficiently.

Feasibility Study

Employee Navigation Training Using 3D Store Mapping

We conducted a study to explore whether AI-assisted indoor navigation could be used as a training tool for new employees in large retail environments.

SEE RESULTS →
3D Store Navigation - Employee Training Tool
Client System · In Production

Turning Messy Emails into Structured Reports

For a moving company, we built a tool that reads unstructured email threads, payment exports, and scanned manifests, then reconciles them into one accurate report — replacing hours of manual analysis and calculation.

SEE RESULTS →
Auto-generated operational report from emails and scanned documents
Feasibility Study

Customer Demographics from Video Analysis

We conducted a feasibility experiment to explore whether existing store camera footage could support high-level demographic analysis without facial recognition or personal data. Click see results below to learn more.

SEE RESULTS →
Video Analytics - Customer Demographics
Feasibility Study

AI-Powered Real Estate Photo Enhancement

We explored whether the latest generative image models could automate professional real estate photo retouching, including window exposure correction and interior enhancement.

SEE RESULTS →
Before - Original Photo After - AI Enhanced Before After
Working Prototype

Smart Recommendations Powered by Your Database

We built an AI-powered search experience for bookstores that turns a simple ISBN inventory into an intelligent, customer-facing discovery tool. Click see results below to learn more.

SEE RESULTS →
AI Book Recommendations

Choosing the Right AI Model

We test AI models on real-world tasks to reveal strengths, limits, and the best fit for your use case.

Vision & Image Analysis

Models capable of interpreting images, identifying objects, and extracting visual meaning.

Models we work with:

  • OpenAI — GPT (image understanding & description)
  • OpenAI — GPT Image (image generation & editing)
  • Google — Gemini (multimodal reasoning)
  • Meta — SAM (object segmentation & tracking)
  • Ultralytics — YOLO (real-time object detection & tracking)
  • OpenCLIP (appearance-based classification)

Audio Analysis & Generation

Models for analyzing, transcribing, separating, and generating audio — from speech recognition to real-time voice conversations.

Models we work with:

  • Meta — SAM Audio (sound segmentation & isolation)
  • OpenAI — Whisper (speech recognition)
  • Nvidia — Parakeet (fast transcription)
  • Google — Gemini Live API (real-time voice conversations)
  • ElevenLabs — Voice cloning & text-to-speech
  • Fish Audio — Voice cloning & speech synthesis
  • Resemble AI — Chatterbox (open-source TTS)
  • Meta — Demucs (music source separation)

Semantic Search & Retrieval

Models optimized for semantic embeddings and similarity search across large document collections.

Models we work with:

  • OpenAI — Text Embeddings (high-accuracy search)
  • Google — Gecko (fast multilingual embeddings)
  • Cohere — Embed (multilingual retrieval)
  • Voyage AI — Voyage (code & technical docs)
  • Hugging Face — BGE (open-source multilingual)

Text & Language Understanding

Models designed to understand, summarize, classify, and reason over written language.

Models we work with:

  • OpenAI — GPT (complex reasoning & analysis)
  • OpenAI — GPT mini models (fast, cost-effective tasks)
  • Anthropic — Claude (long documents & coding)
  • Google — Gemini Flash (speed & multimodal)
  • Meta — Llama (open-source, self-hosted)

OCR & Document Structure

Systems for extracting text and layout from scanned documents, PDFs, and photos.

Models we work with:

  • Google — Document AI (forms & invoices)
  • Microsoft — Azure Document Intelligence (structured extraction)
  • Amazon — Textract (tables & handwriting)
  • Mathpix — Snip (equations & scientific docs)
  • Tesseract (open-source OCR)

Reasoning & Validation

Models used to verify, filter, and sanity-check outputs from other systems.

Models we work with:

  • OpenAI — GPT (complex multi-step reasoning)
  • OpenAI — GPT mini models (fast logical validation)
  • Anthropic — Claude (fact-checking & analysis)
  • Google — Gemini (chain-of-thought)
  • DeepSeek — R models (open-source reasoning)

Many real-world problems require combining several of these capabilities into a single workflow. We help you choose the right models for each step — balancing accuracy, speed, and cost — and test whether the approach holds up before it’s ever deployed.