Full-stack developer and ML builder working where AI, product engineering, real-time software, and local-first tools meet.
INDIA · COMPUTER ENGINEERING · BUILD → SHIP → ITERATE
01 / OBSERVEFind the part of a messy problem that actually costs people time, trust, or clarity. |
02 / MODELTurn the workflow into evidence, states, decisions, and honest constraints. |
03 / ENGINEERConnect product design, full-stack systems, data, and machine intelligence. |
04 / SHIPMake it usable, observable, reversible—and real enough to learn from. |
My build rule: intelligence should expose its evidence, automation should leave the human in control, and ambitious ideas should end as usable products.
01 — Athreix Nexus
Never explain your project to an AI twice. A persistent memory layer that records architecture, decisions, timelines, handoffs, and project knowledge—offline. Explore the memory layer → |
02 — Synapse
A terminal-native control room for AI coding agents. Observe live work, inspect diffs, enforce guardrails, create checkpoints, and rewind changes. Enter the control room → |
|
Research that shows its work. An AI-native platform for evidence-backed company research, buying-signal detection, qualification, strategy, and outreach drafts. Inspect the intelligence engine → |
04 — VYROXX
From athlete evidence to race decisions. Benchmarking, prediction, pacing, causal race replay, and training prescriptions in one mobile-first interface. Open the live product ↗ · Source → |
LANGUAGESRust · TypeScript · JavaScript · Python · Java |
PRODUCTReact · Next.js · Tailwind · responsive interfaces |
SYSTEMSNode.js · Firebase · MongoDB · Redis · Convex |
INTELLIGENCETensorFlow · structured AI workflows · evidence-first UX |
Open the public build signal
If you are building an AI-native product, a developer tool, a real-time experience, or a workflow that needs both systems thinking and a usable interface, let’s talk.
tushar@athreix:~$ build --something-useful
Designed as a system—not assembled as a template.