This blog pairs technical writing with the reasoning behind the work: research reading, engineering decisions, experiments, and project retrospectives. The posts complement repository READMEs with the context, tradeoffs, and revisions that made the work what it is.
5 topics · 19 posts
Writing
Research essays, engineering write-ups, and project stories, grouped by topic.
Showing all 19 posts.
AI Research Deep Dives
8 postsPaper-backed essays
- A Research Program for Inspectable MLAn essay on reading foundational deep learning papers as one research program and turning interpretability and representation analysis into a rigorous science.
- The mental models of deep learningAI research explained through foundational deep learning papers: universal approximation, convolutional architectures, transformers, reinforcement learning, and the open question of task-fit architecture.
- Peeking Inside the Black Box
- Diving Deep into Interpretability
- From Alchemy to Science
- Approaching the Frontiers
- The Era We Live In
- What Intelligence Means for AI Research
Engineering & Research Write-Ups
2 postsProject deep dives
- An Engineering Discipline for Agentic CodingA project write-up on treating agent-assisted development as a verifiable engineering system with operating rules, skills, checks, and review boundaries.
- Forecasting the Limit Order BookAn applied ML write-up on mid-price movement forecasting with logistic regression, MLP, DeepLOB-style sequence modeling, and interpretability probes.
A Developer's Memoir
6 postsProject stories and reflections
- 01Launching Off
- 02Putting Users First
- 03Best Case, Worst Case
- 04Tech for Society
- 05Why AGENTS.md Wasn't Enough for My RepoA first-person account of why scattered instruction files failed me and why a repo-owned boundary became the real problem to solve.
- 06How to Win a HackathonA practical essay on why hackathons are won by matching the idea, demo, and technical work to the organizers and judging criteria.
Beryl in Public
2 postsBuild notes from week one
- 01What Beryl Actually Does (and What It Deliberately Doesn't)A practical walkthrough of the first working structure: `.beryl/agent/`, a deterministic check, and review evidence in files.
- 02Week 1 of Building Beryl in Public: What People Told MeWhat the first feedback cycle exposed in week one: strictness tradeoffs, false positives, and direction for future iterations.
Algorithms Archive
1 postTheoretical computer science essay
Unless otherwise noted, blog illustrations were generated with Stable Diffusion XL under the CreativeML Open RAIL++-M License.