The Job Search Agent: How I’d Add Dedalus to My GTM Job Intelligence Platform

Satvik Panchal — Fri, 10 Apr 2026 23:12:39 +0000

Most job platforms stop at search. They help you find listings, apply filters, and maybe run keyword queries. But job search is not just a retrieval problem. It is a decision-making problem. After building my GTM Job Intelligence Platform, I started thinking about what it would take to turn it from a search tool into a real job search agent.

What I originally built

My GTM Job Intelligence Platform was designed to make job postings easier to analyze at scale. The system:

Scraped 14K+ job postings
Covered 144 companies
Supported 4 ATS platforms: Greenhouse, Lever, Ashby, and SmartRecruiters
Converted unstructured job descriptions into structured fields enabled natural language search, resume matching, and skill-gap analysis

Instead of treating job descriptions as messy text blobs, I wanted to turn them into structured, queryable data.

Tech Stack

Backend: Python, FastAPI
Queueing / Workers: Redis
Database: PostgreSQL
Frontend: Alpine.js, Tailwind CSS
LLM Extraction: GPT-4o
Data Sources: Greenhouse, Lever, Ashby, SmartRecruiters

What the Platform Still Couldn’t Do

Even after building search, filtering, and matching, one major problem remained: the platform could return useful information, but it still left the final decision-making to the user.

A user still had to answer questions like:

Which roles are actually worth applying to?
Which ones are a strong match versus a stretch?
Where am I missing important skills?
Should I apply now or prep first?

That is where I think an agent layer becomes interesting.

How I’d Add Dedalus

I would not use Dedalus to replace the platform I already built. The scraping pipeline, extraction layer, database, and search system would all stay the same. What Dedalus would add is an orchestration layer on top of that foundation, one that turns a data product into a more agentic experience.

Instead of just returning search results, the agent would help users move from information to action. It would understand a user’s goals and constraints, query the existing jobs database, rank opportunities by fit and relevance, explain why certain roles stand out, identify real skill gaps, and recommend what to do next.

That workflow could look like this:

understand the user’s background, goals, and preferences
query the jobs database for relevant opportunities
rank roles by fit, urgency, and overall relevance
explain why each role is a strong or weak match
identify meaningful skill gaps across postings
recommend next steps, such as:
- apply now
- tailor resume first
- study specific topics
- skip for now

That is the shift I find most interesting: moving from a search product that surfaces information to an agentic product that helps users make better decisions.

Why Dedalus + MCP Would Make This Stronger

What makes Dedalus especially interesting here is not just the agent layer, but how MCP could make that layer much safer and more production-ready.

Instead of giving the agent broad access to internal systems, MCP would let me expose only the exact tools and actions it needs. For a product like this, that means tighter control over things like job queries, resume matching, and user-specific recommendations, without opening up raw database access or overly permissive integrations.

That makes the product better in a few ways:

more secure through scoped tool access
more modular as new capabilities get added
easier to maintain than one-off custom integrations
more reliable because the agent works through structured interfaces

That is a big reason why Dedalus stands out to me. It is not just about making the product more agentic. It is about making that agent layer feel controlled, scalable, and realistic to deploy.

The original platform helped answer what jobs are out there.
Adding an agent layer answers the more important question: what should I do next.

That shift — from retrieval to action — is where I think tools like Dedalus actually become meaningful.

DEV Community: Satvik Panchal

The Job Search Agent: How I’d Add Dedalus to My GTM Job Intelligence Platform