Rhizome
★ 2.6kGo · CRDTs · gossip protocol
“Storage that grows like a network — nodes leave, data stays, partitions merge.”
2.6k stars · 1,100 forks · 30% node churn, zero data loss
Builds quiet tools and loud prototypes
I’m Garvit Pandia — a systems & data engineer who builds tools that keep growing in the wild: storage that self-heals, forecasts farmers trust, messages that arrive on bad networks. I studied B.Sc Computer Science at Uttaranga Institute of Applied Sciences, where my independent research on agri-sensor data became the seed of everything since — 8.8/10, and that sensor work still feeds my field projects today. My work spreads along three hyphae: resilient distributed systems (Rhizome survives 30% node churn with zero data loss), environmental data (Kelp warns of acidification six days early), and open source — 176 PRs merged, 31 first-time contributors mentored to their first merge. Wherever the network spreads, I mentor: office hours by terminal, and judging rubrics that teach as they score.
“The acidification warning system he built gave our biologists six days of lead time. Six days is a season in this field.”
“My first merged PR was in his repo. He left 14 comments, all of them kind, all of them right. Now I maintain two projects myself.”
“He took our sensor mess and returned a forecast the agriculture officers actually trust. That is rarer than it sounds.”
“Every chart he built for us carries its parameters. Reproducibility is a personality trait for him.”
“The review culture in Harpoon is the best I’ve seen in any open-source project. He built that with patience, one PR at a time.”
Go · CRDTs · gossip protocol
“Storage that grows like a network — nodes leave, data stays, partitions merge.”
2.6k stars · 1,100 forks · 30% node churn, zero data loss
Python · Next.js · PostgreSQL
“A microbiome dashboard for researchers — every chart carries its exact parameters.”
12k samples in year one · 3 labs · 2 papers cite its export format
Python · PyTorch · Django · PostGIS
“Specimen identification for pressed plants — phone photos in, species and range shifts out.”
93.7% top-3 · 5,800 specimens in 8 months · 2 botany departments
Python · Pandas · D3 · PostGIS
“Hydrology data that tells a story — six feeds, one forecast surface, three river basins.”
87% seasonal accuracy · monsoon onset within 4 days · 12 dashboards
Go · local relays · E2E encryption
“Offline-first messaging that always arrives — phones relay for each other, zero servers.”
99.1% same-day delivery · 610 stars · pilot village on phone relays
Python · scikit-learn · TimescaleDB · Grafana
“Early warning for coastal acidification — buoy streams, six days of lead time.”
91% precision at 48 h · 40k readings/day · 2 research stations
2.6k★ Rhizome — flagship OSS1,100 forks · 30% node-churn survival
176 OSS pull requests merged31 first-time contributors mentored to merge
2 maintained crates on crates.io~1.1M downloads combined
State Innovation Council 2024Best Student Agri-Tech Innovation
9,800+ GitHub starsacross 12 repos
4 npm packages published~320k combined weekly downloads
GitHub Trending, 3× in 2024featured projects across the year
GitHub Star of the WeekRhizome, Sep 2025
Winner — Monsoon Hack Jaipur 2021data-visualization track · 180 teams
Workshop paper — agri-sensor streamsField Data Systems Workshop, 2023
FOSS India — top contributorsfeatured list, 2025
Built the sensor-ingestion path 40+ farm trials run on; caught 3 faulty probe batches before they poisoned a season’s model.
Built the pipeline behind Symbiont — 12k microbiome samples in year one; cut pipeline cost 44% with caching, not hardware.
Flagships Harpoon (3.2k★) and Rhizome (2.6k★); 176 PRs merged, 31 first-time contributors mentored to merge.
Shipped Kelp: pH warnings 6 days early, 91% precision at 48 h; integrated Atlas offline maps for survey teams.
Tuned the Ferry relay for a pilot village — 99.1% same-day delivery; battery-budget guide doubled field-device lifetime.