ESC
Code & AI · Portfolio

Code & AI

Nine open-source computational projects spanning Claude Code / MCP, agentic LangGraph orchestration, CLIP-style PyTorch models, single-cell foundation-model evaluation, and vendor-agnostic lab automation. Drill into one of the three category pages, scroll for the interactive filter, or check the live GitHub activity at the bottom.

Agents & MCP → ML & Foundation Models → Bio Pipelines & Automation → Filter by tech → Live GitHub →
01
02

Filter projects by tech

Click any pill to filter. The tile grid below dims projects that don't match.

All Python TypeScript MCP / Claude Code LangGraph PyTorch Single-cell GitHub Actions Lab automation
03

Live GitHub activity

Real-time view of public commit history, contribution streak, and language mix across every repository on github.com/abayatibrain. Each widget below refreshes from GitHub on page load — no caching, no edits.

Contribution graph — past 12 months
GitHub contribution heatmap for abayatibrain
via ghchart.rshah.org
Commit activity — time series
GitHub commit activity graph for abayatibrain
via github-readme-activity-graph
Overall stats
GitHub overall stats for abayatibrain
Top languages
GitHub top languages for abayatibrain

Widgets are SVG endpoints from 2016rshah/githubchart-api, github-readme-activity-graph, and anuraghazra/github-readme-stats. GitHub blocks iframe embedding of profile pages — this is the next-best thing.

How the suite fits together

01
Wet lab
ipsc-screen-automation produces randomized, QC-clean plates.
02
Imaging
AutoMorphoTrack turns multichannel stacks into quantitative tables.
03
Modeling
morph2expr + scgpt-probe turn features into honestly-evaluated models.
04
Targets
pd-target-credentialing ranks candidates with full provenance and convergent-evidence scoring.
05
Agentic
bioscope's MCP server makes everything callable from Claude Code.