Home/Expertise

Where we work,and what we work with.

Industry context shapes what gets built. The stack is chosen to match the problem, not the other way around. Both are standard tooling your team can maintain after handover.

Start a project See industries

Industries
where the work shows up

Most projects sit at the intersection of a business process and a data problem. These are the contexts we see most often, and the capabilities that usually apply.

Retail and distribution

Forecast demand by SKU and location, optimise inventory, and price against transaction history rather than gut feel.

Subscription and B2B services

Score renewal risk, unify subscription metrics, and give customer success a ranked list each week.

Sales organisations

Prioritise pipeline by likelihood to close, enrich CRM from calls and emails, and forecast from deal behaviour.

Finance and operations

Reconcile billing to operations, predict cash position, and flag transactions or processes that drift from normal.

Manufacturing and logistics

Anticipate failures and delays from sensor data, plan capacity, and detect anomalies before they become losses.

Customer support

Draft replies from your help centre, route by intent, and surface similar resolved cases to whoever picks it up.

Legal, compliance and procurement

Search contracts for clauses and obligations, compare incoming documents to your standards, and extract what matters at volume.

Operations and back office

Process incoming documents, validate against your rules, and post to the system of record with exceptions routed to a person.

The stack
by capability

We favour the simplest tool that meets the requirement. Nothing proprietary, nothing that makes leaving expensive. Model and platform choice follows the problem, not the other way around.

AI models

Model providers and open model infrastructure selected around accuracy, privacy, latency and cost.

  • OpenAI OpenAI
  • Anthropic Anthropic
  • Hugging Face Hugging Face

Machine learning

Data preparation, modelling and experiment tracking for forecasts, scores and classifications.

  • Python Python
  • scikit-learn scikit-learn
  • MLflow MLflow
  • Pandas Pandas

Data pipelines and transformation

Ingestion, orchestration and modelling for reliable, tested data flows.

  • dbt dbt
  • Airflow Airflow
  • Prefect Prefect
  • Airbyte Airbyte

Databases and analytics

Operational databases, warehouses and reporting tools sized to the volume and access pattern.

  • PostgreSQL PostgreSQL
  • BigQuery BigQuery
  • Snowflake Snowflake
  • ClickHouse ClickHouse
  • DuckDB DuckDB
  • Metabase Metabase

Application backends

APIs, application data and caching for production systems that connect models and workflows.

  • FastAPI FastAPI
  • Redis Redis
  • Supabase Supabase

Infrastructure and cloud

Deployment and infrastructure as code on standard platforms your team can maintain.

  • AWS AWS
  • GCP GCP
  • Docker Docker
  • Kubernetes Kubernetes
  • Terraform Terraform
  • GitHub GitHub

Business systems

The systems we connect to, ingest from and write back into.

  • HubSpot HubSpot
  • Salesforce Salesforce
  • Stripe Stripe
  • Slack Slack

Tell us your industry
and what you're trying to fix.

Start a project