Slow & Steady AI: Automating Recruiter Outreach Locally Without Selling My Data

Kashish Rohatgi — Sat, 24 Oct 2026 17:42:08 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady

What I Built
I built a privacy-first, local-running Personalized Email Outreach Agent powered by FastAPI, Webcmd workflow management, and Google’s open-weight Gemma model via Ollama.

Instead of dealing with the exhausting grind of writing dozens of custom cold emails or using risky cloud scrapers that leak professional lists, this agent works methodically and safely: it takes lead data, drafts hyper-personalized recruiter or engineering outreach pitches locally, and routes them strictly into a draft isolation queue. This ensures total data ownership and a mandatory human-in-the-loop review before any message can ever be sent, making the entire networking process effortlessly slow and steady.

Demo
GitHub Repository: KashishRohatgi/personalized-email-outreach-agent

KashishRohatgi / personalized-email-outreach-agent

An AI-driven browser automation tool that generates context-aware, highly customized emails via an LLM and executes end-to-end delivery using webcmd. Built with human-like throttling and a safety draft mode for reliable outreach.

Local Interface: FastAPI interactive Swagger UI available at http://127.0.0.1:8000/docs for testing automated draft generation payloads.

How I Built It
Backend Framework: FastAPI handles execution state, endpoint routing, and payload validation.

Local AI Model: Google Gemma (open-weight) runs entirely offline via Ollama, eliminating external API subscription fees.

Automation & Safety: Integrated with Webcmd execution workflows and structured to isolate outputs inside local markdown files / draft boundaries to maintain strict human-in-the-loop control.

Why Does Open Innovation Matter?
When handling professional networks, target companies, and personal career outreach, data privacy is non-negotiable. Using closed commercial cloud APIs means handing sensitive lead lists and raw communication styles over to third-party tracking servers.

Open innovation via local models like Gemma and open frameworks like FastAPI makes it possible to build powerful productivity automations that run completely offline, ensuring total data sovereignty, zero recurring cloud costs, and complete transparency.

The Lazy Payoff
Manually writing custom cold pitches to dozens of engineering managers or recruiters is a massive time sink that introduces severe decision fatigue. This agent automates the heavy lifting: it batch-processes lead inputs, drafts tailored emails using local AI, and lines them up neatly in a secure review queue. You get to wake up, sip your coffee, review pre-made drafts with one click, and skip the tedious blank-page struggle entirely.
i am applying for gemma competition

DEV Community: Kashish Rohatgi

Slow & Steady AI: Automating Recruiter Outreach Locally Without Selling My Data

KashishRohatgi / personalized-email-outreach-agent

An AI-driven browser automation tool that generates context-aware, highly customized emails via an LLM and executes end-to-end delivery using webcmd. Built with human-like throttling and a safety draft mode for reliable outreach.