🌍 HiringOrbit

Location-Aware Intelligence Agent

Built as part of the GFG x Google for Developers — Build with AI Series (Workshop 2)

🔗 Live Demo: https://hiringorbit.streamlit.app
🔗 GitHub: https://github.com/ankush-poonia007/hiringorbit

---

🧠 What Is HiringOrbit?

HiringOrbit is a location-aware intelligence agent that helps users find relevant places — engineering colleges, tech companies, ML communities, and more — near any city, powered by AI-driven recommendations.

You type a location and a query. The agent finds real nearby places using Google Maps and generates a smart, contextual recommendation using Gemini AI.

---

🏗️ What We Built (Phase 1 — MVP)

The Problem

Finding relevant opportunities (colleges, companies, communities) near you usually means manually searching Google Maps, reading reviews, and piecing together information yourself. There's no intelligent layer that understands why you're searching and gives you a recommendation.

The Solution

A simple but functional AI agent that:

• Takes a natural language query (e.g. "engineering colleges", "tech startups")
• Takes a city/location as input
• Fetches real nearby places using Google Maps + Places API
• Passes that data to Gemini AI which generates a smart, human-readable recommendation

Tech Stack

Layer	Technology	Purpose
AI/LLM	Google Gemini 2.5 Flash	Reasoning & recommendations
Location	Google Maps API + Places API	Fetching real nearby places
Geocoding	Google Geocoding API	Converting city name → coordinates
Frontend	Streamlit	Simple, fast UI
Config	python-dotenv	Secure API key management
Language	Python 3.14	Core logic

Project Structure

hiringorbit/
├── agent.py          # Core logic: Maps API + Gemini AI
├── app.py            # Streamlit UI
├── requirements.txt  # Dependencies
├── .env              # API keys (not pushed to GitHub)
└── .gitignore        # Keeps .env secure

How It Works

User enters city + query
        ↓
Geocoding API converts city → lat/lng coordinates
        ↓
Places API fetches top 5 nearby matching places
        ↓
Gemini AI receives places data + user query
        ↓
Gemini generates a contextual recommendation
        ↓
Streamlit displays results cleanly

---

📸 Screenshots

Homepage

Search Results

Agent Handling No Results

⚙️ How to Run Locally

1. Clone the repo

git clone https://github.com/yourusername/hiringorbit.git
cd hiringorbit

2. Install dependencies

pip install -r requirements.txt

3. Set up API keys

Create a .env file in the root folder:

GEMINI_API_KEY=your_gemini_api_key_here
MAPS_API_KEY=your_google_maps_api_key_here

Get your keys:

• Gemini API Key → aistudio.google.com
• Maps API Key → Google Cloud Console (enable Maps JavaScript API, Places API, Geocoding API)

4. Run the app

python -m streamlit run app.py

---

🔭 What We're Building Next (Phase 2)

Phase 1 is a functional MVP. Phase 2 transforms it into a true agentic system — the original vision from the workshop.

The Upgrade: ADK + MCP Architecture

Component	Phase 1 (Now)	Phase 2 (Planned)
AI Layer	Direct Gemini API call	Google ADK agent with reasoning loop
Tool Access	Hardcoded function calls	MCP servers as pluggable tools
Data	Maps API only	Maps API + BigQuery datasets
Intelligence	Single prompt	Multi-step agent reasoning
Memory	None	Conversation + preference memory
Frontend	Basic Streamlit	Full web app

Phase 2 Features Planned

• ADK Agent — The AI doesn't just respond, it plans: decides which tool to call, in what order, and why
• MCP Servers — Maps and BigQuery exposed as MCP-compliant tool servers the agent can plug into
• BigQuery Integration — Store and query structured datasets: company hiring data, salary ranges, location demographics
• Placement Intelligence — "Find ML companies hiring freshers within 200km of Jaipur with avg salary above 6 LPA"
• Memory Layer — Remember user preferences, past searches, and career goals across sessions
• NeuraLog Integration — Long-term: HiringOrbit's location intelligence layer feeds into NeuraLog, a personal context and memory assistant

---

📌 Why This Project?

This was built as part of a learning sprint to understand:

1. How AI agents connect to real-world data sources
2. How Google's Maps + Gemini APIs work together
3. The foundation of agentic architecture (tool-calling, context passing, reasoning)
4. How to ship a working project end-to-end — not just train models in a notebook

The goal isn't just a demo — it's a foundation for a production-grade career intelligence tool.

---

🛠️ Built With

• Google Gemini API
• Google Maps Platform
• Streamlit
• python-dotenv

---

👤 Author

Ankush Poonia B.Tech AI/ML Engineering — Arya College of Engineering, Jaipur LinkedIn | GitHub

---

Phase 1 complete. Phase 2 in progress.