Runback
Currently building a run orchestration and recovery tool for agentic coding workflows, with execution traces, checkpoints, and replayable state.
Selected builds, experiments, and case studies across agent tooling, data systems, and knowledge work.
Showing all 8 projects
Builds that connect most directly to my writing on systems and tooling.
Currently building a run orchestration and recovery tool for agentic coding workflows, with execution traces, checkpoints, and replayable state.
A Local-first CLI for AI coding experiments, with snapshots, metadata, diffs, evaluation results, and restore workflows.
A Telegram-driven review loop where Hermes grades answers against an Obsidian wiki corpus, tracks mastery in SQLite, and schedules follow-ups with SM-2-inspired intervals and category-targeted grill sessions.
A Chrome extension for bookmarking your reading position in long AI chat threads, with stable anchors, scroll restore, and adapters for ChatGPT and Claude.
An end-to-end Airflow ETL pipeline that ingests social and news posts into BigQuery for sentiment analytics.
Smaller builds and older experiments that round out the portfolio.
A knowledge management platform with backend APIs, PostgreSQL schema work, and enrichment pipelines for faster retrieval.
A deal aggregation tool that standardizes noisy listing data for search, comparison, ranking, filtering, and refresh workflows.