AI Agent Developer / Vibe Coder — Full Stack

AI Agent Developer / Vibe Coder — Full Stack

For Learn with Leaders.

Why this role exists

We're not "adding AI to our workflows." We're rebuilding the company as an AI-native org — every internal tool, every customer touchpoint, every dashboard gets shipped by a small squad using modern AI-powered dev tools. You're the person who turns

"we need a thing by Friday"

into a working product by Wednesday.

What you'll actually do

Ship a tool a day.

That's the cadence. Scrapers, research agents, data enrichment flows, internal dashboards, one-off automations. Some will live forever, some will die in a week. Both outcomes are fine. What's not fine is spending a week on "architecture." Build multi-agent systems that actually do work.

Research → enrich → qualify → draft → schedule. Not a single clever prompt — orchestrated flows with agents that hand off to each other, retry cleanly, and log what they did. Scrape everything.

Schools, decision-makers, programs, competitors, grant databases. Firecrawl, Apify, Playwright, BeautifulSoup — whatever works. Feed the pipeline. Keep it fresh. Vibe-code full-stack apps.

Lovable, Antigravity, Cursor, v0, Bolt, Claude Code — pick the right tool per job. Idea → deployed app in a day, not a sprint. Own n8n.

Scale the workflow backbone. Pair with Sujit; the library doubles on your watch. Every broken workflow gets fixed same-day. Integrate LLMs like an adult.

Prompt engineering, tool use, RAG, evals, guardrails, cost monitoring. You know the difference between a flaky prototype and a reliable agent — and you ship the reliable one. Pair, document, ship.

Work with Aditya (AI Automation) and Anisha (Business AI Analyst). Every agent has an owner. Every workflow has a README.

Tech you'll live in

Vibe-coding & agentic IDEs:

Lovable, Antigravity, Cursor, Claude Code, Windsurf, v0, Bolt

Automation:

n8n (primary), Zapier, Make

Agents & LLMs:

Claude API, OpenAI, LangGraph, CrewAI, MCP servers, vector DBs (Supabase, Pinecone, Qdrant)

Scraping & data:

Firecrawl, Apify, Playwright, BeautifulSoup, Scrapy

Full stack (non-negotiable):

Next.js, React, Node.js, Python, Postgres / Supabase, Vercel, auth, webhooks, APIs

Nice to have:

Dockerfile you've actually written, basic cloud (AWS/GCP), observability (Langfuse, Helicone)

You'll be great at this if

You have a

portfolio of shipped things

— links, screenshots, GitHub. We don't care about degrees; we care about what's live. You have

real full-stack chops.

Vibe-coding is an accelerant, not a crutch. When Lovable generates broken code, you fix it. You can read and debug a Postgres query, a React hook, and an n8n node with equal ease. You

ship daily — plural.

Tool-a-day isn't a KPI here, it's your natural speed. A week with nothing shipped feels wrong to you. You've built

multi-agent flows , not just single-prompt wrappers. You know when to use an agent, when to use a workflow, and when to just write a Python script. You've

scraped the internet

— and know which sites fight back and how to handle it. You're genuinely curious. You've already tried the thing that dropped last week. You have opinions on Antigravity vs. Cursor vs. Claude Code. You can talk to non-technical teammates without making them feel stupid.

Bonus points

Built something with schools / edtech / nonprofit ops in the loop Open source contributions Strong design sense — your UIs don't look like it's 1997 You have opinions on framework choices (and can defend them without being annoying)

What success looks like — first 90 days

No shadowing. No ramp-up quarter. You ship on Day 1. TimelineOutcome Day 1 One working thing shipped by EOD. Could be a scraper, a small agent, a Lovable internal tool. Doesn't have to be pretty. Has to be live. Week 1 5+ tools/agents/flows in production. Plugged into n8n. At least one connects Lovable → n8n → our DB. Day 30 ~20 shipped artifacts.

2 multi-agent systems

live (e.g., research → enrich → qualify → outreach draft). One Lovable-built internal dashboard that a non-tech teammate uses daily. Day 60 You own a full surface end-to-end — could be an agentic sales workflow, a scraping + enrichment pipeline feeding the CRM, or a customer-facing Lovable app. Tool-a-day is your steady state, not a spike. Day 90 The AI squad's throughput has 3x'd because of you. Nothing you've shipped has been rebuilt. The team genuinely can't imagine working without you. The bar:

if your instinct on Day 1 is "let me set up my dev environment properly first," this isn't the role. The instinct should be "what's the smallest useful thing I can deploy in the next 4 hours?" What we offer

Competitive comp + meaningful equity / bonus structure Every AI tool on us

— Cursor Pro, Lovable Pro, Antigravity, Claude Max, n8n Cloud, Firecrawl, Apify — if it helps you ship, we pay for it. Autonomy.

You ship. We don't do stand-ups as performance art. Direct visibility to the CEO via the CIO. A small, senior team. No committees. No "let's circle back." Work that actually reaches schools and students across India and Africa.

How to apply Send us three things: Your 3 favorite shipped projects

(links — not descriptions) A 3-minute Loom

walking us through one multi-agent flow you've built (what it does, which tools, what broke, how you fixed it) GitHub / portfolio No cover letters. No CV templates. Just show us what you've built.

The working interview If we like your application, we send you

one brief and 48 hours.

Something like:

"Scrape this list of 500 schools, enrich with decision-maker emails, draft personalized outreach via Claude, push to n8n, deploy a Lovable dashboard to monitor replies."

You ship it. We review. That's the interview. No whiteboarding. No LeetCode. No "tell me about a time."

If you've been building AI agents in your bedroom for a year and have nothing "official" to show for it but the work is sharp — apply anyway. We'll read it. Please email ipsita@learnwithleaders.com.