Satvik Adyanthaya — an archive in five rooms

Rochester · NY 43.16°N 77.61°W · MMXXVI

“The cost of being wrong, multiplied by the probability of being wrong, is the only number that matters.” — S. A. “All models are wrong, but some are useful.” — George E. P. Box “I'd rather build systems you can run than diagrams you have to defend.” — S. A. “Plans are worthless, but planning is everything.” — Dwight D. Eisenhower “Most decisions are reversible. The interesting work is at the few that aren't.” — S. A. “Invert, always invert.” — Carl Jacobi “An agent that asks for help beats an agent that tries harder.” — S. A. “I would rather have questions that can't be answered than answers that can't be questioned.” — Richard Feynman “Without the critique step you don't have a pipeline — you have a generator.” — S. A. “Premature optimization is the root of all evil.” — Donald Knuth “Every model is a mortgage on attention. Choose the ones you can afford.” — S. A. “The best way to predict the future is to invent it.” — Alan Kay “Strategy is whatever survives a four-vertical product review.” — S. A. “We are drowning in information and starving for wisdom.” — E. O. Wilson “The good decks are written for the year after the meeting, not the meeting itself.” — S. A. “What can be asserted without evidence can be dismissed without evidence.” — Christopher Hitchens “Forecasting is just memory pointed forward.” — S. A. “Make things as simple as possible, but no simpler.” — Albert Einstein “The hard part of multi-agent systems is deciding which agent gets to be wrong this round.” — S. A. “He who has a why to live can bear almost any how.” — Friedrich Nietzsche “A hypothesis you cannot simulate is just a hope.” — S. A. “The first principle is that you must not fool yourself — and you are the easiest person to fool.” — Richard Feynman “I want to build the boring half of intelligence — the part that remembers, accounts, and apologizes.” — S. A. “The map is not the territory.” — Alfred Korzybski

↻ drag · click a node

humanovo — biomedical discovery, 2026.

Strategic Foresight — scenario planning, 2025.

ArchDisc — AI design platform, 2025.

Sports Decision Intelligence — front-office analytics, 2025.

GroveCart — e-commerce, founded, 2023–24.

satvik.me — this monograph, 2026.

01 / 06 2026

humanovo

biomedical discovery

An AI-native platform for biomedical discovery. A team of specialised agents reasons over a hybrid memory of biomedical literature and clinical data to surface, formalise, and stress-test research hypotheses before a human ever sees them.

Founder, Architect Concept Award · NYBPC FLX 2026

02 / 06 2025

Strategic Foresight

scenario planning

Enterprise scenario planning across defense, energy, healthcare, and finance. A multi-agent system with a smart routing layer places every piece of work on the most cost-effective model — high-end forecasts at a 40% reduction in inference spend.

Architect, Lead Engineer Shipped · 4 verticals

03 / 06 2025

ArchDisc

AI design platform

Procedural geometry engine generating 350+ architectural components from natural language. Eight workbenches, real-time WebGL rendering, single-canvas CAD/BIM/VFX continuity.

Founder, Builder MVP · paused

04 / 06 2025

Sports Decision Intelligence

front-office analytics

A Bloomberg terminal for football clubs — real-time squad valuation, FFP compliance, RAG-powered scouting across performance and transfer data.

Founder, Builder MVP · paused

05 / 06 2023–24

GroveCart

e-commerce, founded

Founded and shipped solo. 1,000+ SKUs, AI checkout chatbot, full payments stack. 20% MoM growth before handoff.

Founder, Sole Engineer Shipped · live

06 / 06 2026

satvik.me

this monograph

A six-page editorial on what I'm trying to build. Astro 5, scroll-snap deck, hand-tuned typography, three.js knowledge graph. The thing you're reading.

Author live · this site

§ drag the orb · click a node

§ stack

Languages

Python TypeScript SQL R JavaScript Bash

AI · ML

PyTorch LangChain LangGraph Claude Gemini AWS Bedrock SageMaker RAG multi-agent orchestration embeddings fine-tuning Hugging Face

Data

Pandas NumPy Scikit-learn Monte Carlo A/B testing forecasting Tableau Power BI

Backend

FastAPI Node.js Next.js PostgreSQL Neo4j ChromaDB MongoDB Redis

Cloud

AWS GCP Vercel Docker Kubernetes CI/CD

Biomedical

genomic pipelines UMLS / MeSH PubMed mining clinical-trial data biomedical NLP

§ tongues

English fluent
Hindi fluent
Tulu native
Spanish conversational
Marathi conversational

§ certified

Google Project Management 2024
Industrial Biotechnology 2023
AWS Cloud Practitioner 2025

§ roles

VP, Simon Entrepreneurship Association 2025– U of R
Merit Scholarship 2025 Simon · Rochester
Founder, GroveCart 2023–24 Mumbai

§ learning

graph neural nets for drug discoveryBayesian decision theoryPyMC for hierarchical modelsWebGPU compute pipelinesMandarin

§ first principles

Optimise for being wrong cheaply, not for being right loudly.
Most decisions are reversible — spend judgement on the few that aren't.
An agent that knows when to ask beats an agent that tries harder.
Build the boring half of intelligence: memory, accounts, apologies.

§ on the desk

Thinking, Fast and Slow Kahneman
Antifragile Taleb
The Signal and the Noise Silver
Superforecasting Tetlock

§ what I reach for

2025 — 2026

M.S. Business Analytics
University of Rochester · Simon
Rochester, NY
Sep 2025 — present

Technical Assistant
University IT, U of R
Rochester, NY
Jan — Jun 2025

Project Lead, R&D
Organic Recycling Systems
Mumbai
Jun — Jul 2024

Research Intern, Drug Development
Glenmark Pharmaceuticals
Mumbai
Oct 2023 — Apr 2024

Founder & Developer
GroveCart
Mumbai
2021 — 2025

B.Tech Biotechnology
Amity University
Mumbai

§ what got me here

write to satvikconnects@gmail.com

|

§ stack

Languages

AI · ML

Data

Backend

Cloud

Biomedical

§ tongues

§ certified

§ roles

§ learning

§ first principles

§ on the desk

M.S. Business Analytics

Technical Assistant

Project Lead, R&D

Research Intern, Drug Development

Founder & Developer

B.Tech Biotechnology