PredictAndProfit.io
A self-hosted Kalshi trading suite and its public website: release qualification, exchange reconciliation, a forecast-model audit, and sitemap engineering, with a source for every figure.
Read the case studyCase studies
Focused examples of turning ideas, workflows, and technical requirements into software that can be owned and extended.
Technical field notes
Client case studies need permission before the numbers can be published, and that takes time. In the meantime these are write-ups of real technical work, in enough detail to judge whether we know what we are doing.
Workday SOAP APIs, custom connectors, and rate limits
Building against a large enterprise SOAP API without turning the sync into a slow-motion incident: pagination, incremental state, payload size, and retry design.
Read the write-upCxAlloy API extraction guide: request signing, nested pagination, rate limits and reconciliation
A runnable extraction of nested CxAlloy listings that fails mid-run, resumes at the failed page within an hourly request budget, and reconciles every record on a composite key.
Read the write-upPipeline alerting reference guide: AWS Lambda, DynamoDB, Teams and Jira without duplicate incidents
A runnable alerting system that opens one incident per failure, suppresses repeats with DynamoDB conditional writes, survives Teams and Jira outages and a crash, and reconciles what was sent.
Read the write-upFivetran Connector SDK production guide: state, retries and restart safety
A runnable connector that fails mid-sync, restarts from its last checkpoint, and reconciles every row by primary key, with the Fivetran behaviors it relies on cited.
Read the write-up