Languages
- TypeScript
- Python
- Go
- Daml
00 / Atharva Ashtekar / 2026
--:--:-- IST
n8n verified creator
Available for new work
AtharvaAshtekar
- 01AI / ML Engineer
- 02Automation & Agents
- 03Product Engineer
- 04Founder
I build AI systems, automation and the products around them: agent and workflow pipelines, the ledger underneath, and the interface people use.
Currently building
See the work- TenzroDistributed ledger infrastructure and self-custody wallets on Canton.
- AI AutomationsAgent and workflow pipelines in n8n and TypeScript that remove the manual steps between systems.
- TypeScript
- Python
- Go
- Daml
- LLM agents
- RAG / ChromaDB
- Claude API
- Evals
- Prompt architecture
- n8n
- Workflow automation
- Webhooks
- API integrations
- React 19
- Next.js 15
- Tailwind CSS
- FastAPI
- PostgreSQL
- Stripe
- Turborepo
- Cloudflare
- gRPC
- Playwright
- Three.js / XR
~ / Toolkit
What I build with.
01 / Selected work
SomethingsIbuiltandshipped.
Ledger infrastructure, multi-tenant AI SaaS, applied research tooling and agentic pipelines. Each one started as an empty repository and runs today.
01 / Distributed ledger infrastructure / 2025 to now
Tenzro Network
Ledger infrastructure and a self-custody wallet built on the Canton privacy-preserving network.
Network services and wallet clients around a privacy-preserving distributed ledger: key management, transaction construction and settlement flows, with the contract layer written in Daml so that who-sees-what is enforced by the ledger rather than by an application check.
- Role
- Founder / Engineer
- Stack
- Daml / Canton / TypeScript / Node.js
Automations
02Workflow automation suite / 2025 to now
Agent and workflow pipelines in n8n and TypeScript that remove the manual steps between systems.
- n8n
- TypeScript
- LLM APIs
- PostgreSQL
Automation Engineern8n profile →
Kolektiva
03RWA fractional property platform / 2024 to 2025
A real-world asset platform that tokenises property for fractional on-chain ownership.
- Solidity
- Ethereum
- Web3.js
- Next.js
Engineer
Kolfly
04Multi-tenant AI SaaS / 2025
A multi-workspace, AI-powered B2B platform for influencer marketing, from discovery through to billing.
- Next.js
- TypeScript
- LLM pipelines
- PostgreSQL
Product Engineer
AI Tutor
05Applied AI research artifact / 2025
An LLM classroom tutor with three avatar modes, built as the usable artifact for a research study.
- Python
- LLM orchestration
- Real-time avatars
- WebRTC
Engineer
AI Website Redesign Pipeline
06Agentic tooling / 2025
A human-in-the-loop pipeline that scrapes a site, generates redesigns, and exports a real Next.js project.
- Next.js 15
- Playwright
- Claude API
- TypeScript
Founder / Engineer
WMTI
07Polyglot platform / 2026
A Turborepo monorepo running three services in three languages behind one shared type contract.
- Turborepo
- React 19
- Go / Fiber / gRPC
- FastAPI
Architect / Engineer
Canton Experiments
08Open pattern library / 2025
A library of self-contained Daml contracts, each isolating one distributed-ledger pattern.
- Daml
- Canton 3
- Daml-LF 2.x
Author
02 / What I do
WhatIgethiredtodo.
Four lanes, one person. Most engagements start in one and end up crossing all four.
A01
AI systems
Agents, retrieval and LLM pipelines built to survive contact with real inputs: evaluated, bounded, and cheap enough to run in production rather than in a demo.
- Agents
- RAG
- Evals
- Claude API
- Prompt architecture
B02
Product engineering
The whole product, not a layer of it. Data model, API, front end and deploy, shipped as one coherent thing by one person who holds the entire picture.
- Next.js
- React
- Go
- FastAPI
- PostgreSQL
C03
Automation
n8n verified creatorAgent and workflow automation, mostly in n8n and TypeScript: webhooks, scheduled jobs, API integrations and approvals. I am an n8n verified creator, so the workflows run on a platform I know end to end.
- n8n
- Agents
- Webhooks
- API integrations
- Human in the loop
D04
Ledger & Web3
Privacy-preserving distributed ledger work in Daml on Canton. Contracts where confidentiality and settlement are properties of the ledger, not of the app on top.
- Daml
- Canton
- Smart contracts
- Wallets
E05
Zero to one
Taking an idea from a paragraph to a running system with users on it, including the unglamorous middle where most prototypes quietly die.
- Architecture
- Prototyping
- Founding engineering
03 / Process
Howaprojectmoves.
Four stages. Automate sits in the middle because that is where most of the time gets saved.
01
Understand
- Requirements
- Constraints
- System design
02
Automate
- APIs
- n8n
- Agents
- Webhooks
03
Build
- Next.js
- TypeScript
- Backend
- Integrations
04
Ship
- Deploy
- Monitor
- Iterate
04 / About
Oneperson,bothendsofthestack.
I build the whole product, not one layer of it. Most of what I ship starts as a rough idea and ends as a running system: the retrieval and agent layer, the API and data model under it, deployment, and the front end people use.
01
Own both ends
The model and the data, the API and the interface. Keeping both ends in one pair of hands stops the seams from becoming someone else's problem.
02
Automate the middle
The manual steps between systems are where time leaks. Agent and workflow pipelines remove them, with a human left on the calls that matter.
03
Build for real users
A prototype is the start, not the finish. I build the version that survives real inputs, real load and real people.
- Products carried from an idea to production
- Zero to oneProducts carried from an idea to production
- Languages running in production
- PolyglotLanguages running in production
- Research artifact behind a published study
- PublishedResearch artifact behind a published study
- Full-stack ownership
- SoloFull-stack ownership
Currently / Manexus · Tenzro
Two ways in. Book a 30-minute slot, or send a short brief and I reply by email.
- GitHub@AthAsh007
- LinkedInin/atharva-ashtekar
- n8n@athash007
- WhatsApp+91 95189 70722
- Telegram@DeadAss007
- Emailathatharva2002@gmail.com
Available for new work · India, working remotely