When the system you need does not exist yet.

Web and data acquisition, analytical engines, automation pipelines and integrations, engineered for a problem with no off-the-shelf answer.

Some problems stop exactly where off-the-shelf tools end. We build the system that picks up there: web and data acquisition, an analytical engine, an automation pipeline or an integration, purpose-built around your exact problem and fully yours to keep. This is not theory. It is the same engineering behind four live products built from scratch, The 13F, Souppe, SentiSift and Capitol Gains, by someone who pairs 12+ years in institutional markets with 7+ years shipping production systems in Python (FDP), Tom Pickel. What was built for those products gets built for your problem.

Who this is for

  • You need web or data acquisition from sources with no clean API
  • You need an analytical engine: scoring, ranking, matching or a model in production
  • You need a pipeline or an integration between systems that do not talk
  • Off-the-shelf tools stop where your problem starts

What you get

  • A purpose-built system engineered around your exact problem
  • Built in Python, delivered ready to run and documented
  • Yours to keep and run, with no lock-in
  • The same engineering proven in four live products

How it works

  • Describe the problem and what a good outcome looks like
  • We scope the smallest system that solves it
  • You receive a working, documented system you own

Common questions

What have you built this way?

Four live products from scratch: The 13F (intelligence from SEC filings across 8,600+ managers), Souppe (an eight-dimension portfolio scoring and recommendation engine), SentiSift (a real-time moderation API using an ensemble of 8+ AI models at 94% accuracy) and Capitol Gains (congressional trade alerts from official US House disclosures).

Can you acquire data from a source with no clean API?

Yes. Web and data acquisition is core to the work. That is how The 13F reads SEC filings and how Capitol Gains reads official US House disclosures.

Can you build AI or NLP into it?

Yes. SentiSift runs an ensemble of 8+ AI models in production at 94% accuracy across languages. The same approach applies to your problem when it fits.

Do I own the system?

Yes. It is built in Python, delivered ready to run, fully documented and yours to keep, with no lock-in.

Related work: the products this engineering built, SEC 13F intelligence (The 13F), portfolio risk analysis (Souppe) and comment intelligence (SentiSift). For recurring market and portfolio data, see investment data systems; for back-office automation, see operational data and process systems. See the full advisory practice.

Bring the problem with no off-the-shelf answer.

Send a short note describing the problem and what a good outcome looks like. You get a clear next step and a plan for a system you own.

Start a conversation