Services
DF1 takes on work where the analysis and the system that delivers it have to be right together. Each service below is backed by a system already running.
Data analytics and modeling
Econometric and statistical analysis of a defined question: panel models, event studies, forecasting, and scenario analysis. Each result comes with the assumptions that produce it and the limits within which it holds.
Machine learning and AI systems
Predictive models and retrieval-augmented systems for structured and unstructured data. Models are validated against data they have not seen, and language-model output is grounded in source text a reviewer can check.
Data engineering
Migrations, pipelines, and databases built so that their correctness is tested rather than assumed: dependency ordering, type handling, and source-to-target parity checks.
Systems and workflow design
The applications and operating processes that carry an analysis into daily use: APIs separated from interfaces so the numbers can be tested, ingestion that keeps data current, and review steps placed where errors are expensive.
Domains
The methods are general; the depth is not. DF1 works most closely in energy and infrastructure, where grid capacity has become the constraint on where new compute can be built; in finance and markets, from valuation to risk and volatility; and in document-heavy domains, where a confident wrong answer costs more than no answer.
Delivery
An analysis that stays in a notebook does not change a decision. Every engagement ends with something usable: a documented model, a deployed application, or both. The systems listed underwork are running and open to inspection.
Engagement shapes
- Analysis
- A defined question answered with a documented model and a written result.
- Build
- A model or pipeline delivered as a deployed application your team can use.
- Advisory
- Ongoing review of method, data, and interpretation for an in-house team.