GenAI

~

LLMs

~

RAG

~

OpenAI

~

Claude

~

Gemini

~

YOLO

~

SAM

~

MCP

~

MLOps

~

Embeddings

~

ViT

~

ModernBERT

~

FAISS

~

PyTorch

~

CoreML

~

GenAI ~ LLMs ~ RAG ~ OpenAI ~ Claude ~ Gemini ~ YOLO ~ SAM ~ MCP ~ MLOps ~ Embeddings ~ ViT ~ ModernBERT ~ FAISS ~ PyTorch ~ CoreML ~

We turn messy data and AI hype into working solutions. Whether you’re scaling fast or just getting started, work with us to get results.

Who We Are

Loom Labs is Dilpreet and Gagan, with a few other hands on deck. We've been building production ML systems for 10+ years, way before "LLMs" became the only thing anyone wanted to talk about. We've worked everywhere from 3-person startups to universities to government projects. We've built quick prototypes and high-throughput systems that have processed over a billion inferences.

We're small, so there are no sales calls - just honest technical conversations about whether we're the right fit for your problem.

More about us →

How We Work

Fixed scope. Fixed price. Fixed timeline.

Before we start, you'll know exactly what you're getting, what it costs, and when it'll be done. No ambiguity and no surprises.

Here's what that looks like:

Discovery & scoping: We start with a simple conversation. We'll dig into your problem, data, constraints, and goals. If we're not the right fit, we'll tell you.
Statement of work: You’ll get a clear proposal of deliverables, milestones, and a fixed price.
Regular check-ins: We work in short cycles with frequent demos. If something needs to change, better to do it early.
Clean handoff: Every project ends with documentation, a walkthrough, and (if you need) training for your team. We don’t like lingering projects, and likely, neither do you.

We’ve Worked With

We've worked with everyone from early-stage startups to universities and large organisations. Here are a few.

What We’ve Delivered

Generative Terrain Shadings

Eduard integrates in-house developed generative models to help cartographers create beautiful shaded reliefs for terrain maps. A web version is in active development for wider accessibility.
Anomaly Detection

A suite of vision models to help process over 25 million high-resolution transmission line imagery; extracting over 98 million features; detecting conductor issues with an accuracy rate > 99.5%. See more at Raedyne Systems
Automated Document Extraction using LLMs

Production LLM pipeline for extracting key information from highly technical PDF documents. Evaluated on 8,000+ pages, deployed to process millions.
Docket - Expense Management App

LLM powered expense tracking app that uses both on-device & cloud AI to understand and extract key details from documents. Built, launched, and maintained on the Apple App Store. Learn more
Building GenAI Evaluation Loops

Dilpreet's talk at PyCon Australia 2025 on building evaluation loops that actually work when traditional metrics fail (See more)
Spatial Visualisation of Historic Maps

Globes is an immersive spatial experience for the Apple Vision Pro, created with Monash University and Stanford's David Rumsey Map Center.

Work With Us

We’re not here to upsell you on a platform or lock you into an endless retainer.

We help you ship AI and data systems that actually work. Whether you're building your first LLM tool, fixing a broken data pipeline, or trying to figure out which AI projects are worth doing, we'll help you cut through the noise and get to working code.

We're probably a good fit if

You're a non-technical founder or leader trying to figure out if AI can actually solve a real problem in your business
You've got messy data sitting in spreadsheets, databases, or PDFs and need to make sense of it
You want to know if an LLM can handle a specific task before building a whole product around it
You need computer vision working in production, not just a demo
You're tired of consultants who bill by the hour with no incentive to finish
You want someone who'll tell you if AI isn't the right solution

We're probably not a good fit if

You just want a chatbot wrapper around ChatGPT (check out Upwork)
You're looking for the cheapest/quickest option (quality evaluation + models take time)
Your timeline is "we need this yesterday"

What it costs

Most projects range from $15K-$50K depending on scope and timeline. You'll know the exact price before we start

We’ve helped startups, scale-ups, and teams inside larger orgs make sense of AI.

Get in touch.

Contact

Whether you’re leading a team or just trying to push things in the right direction from the inside, we’re happy to chat. We offer an initial free consultation to anyone curious enough to reach out.

Schedule a quick call: Calendly

If you prefer email: info@loomlabs.au