Log 001 — Surface

Garvit
Pandia

I make machines legible

Descending through intelligent systems, interactive worlds, and the occasional trench of production code. Bioluminescence included.

14buoys monitored
40kreadings per day
6 daysof early warning
310kdemo plays
96%positive reviews
99.98%uptime · 14-month record
Begin the dive ↓ Download resume

Est. descent — 11,204 m to the floor · Oxygen: infinite · Curiosity: critical

Log 002 — Diver profile

About the diver

Depth 2,240 m — Twilight Zone

Garvit Pandia is an engineer of quiet systems, based in Kochi, Kerala. The projects here are field notes, not exhibits — some are still growing. He descends through the full stack, from sensor streams to renderers, and brings working specimens back to the surface.

Education: B.Tech Computer Science & Engineering, Verghese Institute of Technology, Kochi · 2020 – 2024 · CGPA 8.2/10 · founded the campus open-source society (40 members by graduation). Current depth rating: comfortable at 11,204 metres of production code.

He ships at depth: 176 pull requests merged across 12 open-source repos, 46 projects shipped end-to-end, and a conviction that the best interfaces feel alive — like something glowing in the dark.

Mission manifest

RoleI make machines legible
BaseKochi, Kerala
EducationB.Tech CSE — Verghese Institute, Kochi
Tenure2020 – 2024
StandingCGPA 8.2/10 · OSS society founder
Signalgarvit9829@gmail.com
Max depth11,204 m — Challenger Deep

Signal log — surface reports

"The acidification warning system he built gave our biologists six days of lead time. Six days is a season in this field."

— Kavya Menon, Marine Data Director, Peregrine Deep

"He found three bugs in our data pipeline that would have surfaced during harvest. The harvest thanked him."

— Shreya Pillai, UX Researcher, Velora Systems

"He beta-tested our field spectrometer in conditions we wouldn't ship a phone into. The calibration log he sent back was better than our QA's."

— Devendra Pillai, Lab Head, Pebblebrook Instruments

Log 003 — Specimens

Specimens collected

Depth 4,480 m — Midnight Zone

Specimen 01 — 9,120 m

Kelp

"Early-warning system for coastal acidification — pH anomalies forecast up to 6 days ahead from buoy streams."

91% precision at 48 h · 40k readings/day · 14 buoys

Pythonscikit-learnTimescaleDBGrafana

Specimen 02 — 7,990 m

Kelpfield

"Kelp-forest regrowth tracked from space and sonar — 11 restoration sites with confidence bands."

11 sites tracked · 92% confidence · 3 NGOs

PythonSonar fusionDashboards

Specimen 03 — 6,400 m

Tidewalker

"An underwater exploration game — a 2 km² reef generated from real depth data."

310k demo plays · 60 fps mid-range · 11 species

three.jsWebAudioProcedural reefs

Specimen 04 — 5,700 m

Stratum

"The ground, visualized — borehole cross-sections in one command, drawn entirely from data."

400+ logs in 2 min · 480 stars

PythonGDALPostGIS

Specimen 05 — 8,350 m

Shortwave

"Voice and field-audio streaming that survives bad links — down to 4 kbps, still intelligible."

94% intelligibility at 40% loss · 340 stars

C++OpusJitter buffers

Specimen 06 — 4,800 m

Dipole

"A $25 SDR dongle becomes a waterfall scanner of three bands — signal classification built in."

11,000+ hours logged · 91% accuracy

C++SDRWebAudio

Specimen 07 — 9,940 m

Warpaint

"26 procedural brushes with seeded randomness — same seed, same stroke, forever."

120 fps with 800-brush strokes · 45k downloads

TypeScriptCanvas 2DWebGL

Specimen 08 — 3,200 m

Cusp

"Magnetosphere data for non-specialists — live aurora-forecast maps from 15 merged sources."

2,800 daily users during solar max · 82% correct calls

PythonD3Satellite feeds

Log 004 — Crew

Bioluminescent crew

Depth 6,720 m — Abyssal Zone

Languages

TypeScript
94
Python
93
C++
72
Rust
84
Go
78
SQL
88

AI / ML

PyTorch
86
Edge inference
84
Time-series forecasting
86
Retrieval systems
88
Evaluation harnesses
90
Model compression
80

Web / Creative

React
93
Next.js
88
Three.js
85
WebGL
80
Canvas 2D
90
CSS motion & animation
94

Systems / Cloud

Linux
88
Docker
86
Kubernetes
78
PostgreSQL
90
Redis
84
AWS
82

Log 005 — Depth records

Depth records

Depth 8,960 m — Deep Hadal

Depth 8,848 m · 1,086 atm

Top 2% — Ocean Temperature Forecasting 2024

global leaderboard · 2024

Depth 10,102 m · 1,240 atm

Paper — Low-Overhead Tracing

Journal of Open Tools & Systems · 2025, peer-reviewed

Depth 7,421 m · 911 atm

Panelist — EdgeCon India 2025

edge AI in the field

Depth 5,486 m · 673 atm

2 regional design honors

Gilded client sites · 2025

Depth 6,401 m · 786 atm

Winner — Monsoon Hack Jaipur 2021

data-visualization track · 180 teams

Depth 9,204 m · 1,130 atm

2 maintained crates

~1.1M downloads on crates.io

Depth 11,034 m · 1,354 atm

9,800+ GitHub stars across 12 repos

combined · 2026

Log 006 — Descent timeline

Descent timeline

Depth 10,080 m — Hadal Plain

Feb 2026 – present · Peregrine Deep

Research Engineer

"Leading the sensor-fusion model for 14 coastal buoys (40k readings/day); designing the open dataset release — every number traceable to a public source."

Apr 2025 – Jan 2026 · Peregrine Deep

Data Lead (contract)

"Shipped Kelp — pH anomaly warnings 6 days early, 91% precision at 48 h; built the Kelpfield monitoring dashboard for 3 restoration NGOs."

Apr – Jun 2025 · Larkspur Bio

Bioinformatics Contractor

"Ported the Canopy drone pipeline to a 1-GPU footprint — 40 km² mapped, 1.1M trees; published the export format, now cited by 2 papers."

Log 007 — Resurface

Resurface

Depth 11,204 m — Hadal Zone · The Floor

Hadal zone — the bottom, where you resurface

Emailgarvit9829@gmail.com Phone+91-8905402023 GitHubgithub.com/garvit-pandia LinkedInlinkedin.com/in/garvit-pandia ResumeGarvit_Pandia_Resume.pdf