Hermes agent automation engineer/ Ai agent builder

We are looking for a highly resourceful AI automation engineer to configure and build a working Hermes Agent–based AI operations assistant for our company. This is a short, execution-focused project. We already know what we want the agent to do. Your responsibility is to configure Hermes, connect the required tools, build workflows, create prompts/skills, test everything, and deliver a usable system within 3–4 days. You should already be comfortable with: * Hermes Agent * AI agents / LLMs * APIs * MCP servers * Browser automation * n8n / Make / Zapier * Python / JavaScript * Gmail / Google Workspace APIs * Slack / messaging integrations * AI model routing * Cron jobs / scheduled automation * VPS / Linux environments You must be someone who can figure things out independently rather than requiring step-by-step instructions. ⸻ What We Want Built The goal is to turn Hermes into an AI-powered Founder Operations / Research / Automation Assistant. The system should be able to handle as much of the following as technically possible. 1. Research & Intelligence Hermes should be able to: * Perform detailed web research * Research software, SaaS companies, agencies, vendors, and service providers * Compare pricing, features, limitations, pros, and cons * Research competitors and market opportunities * Search documentation * Research Reddit, forums, reviews, and community feedback * Verify information using multiple sources * Produce concise recommendation reports * Save important research for future reference Example: Find 10 virtual-card providers that accept UAE companies, compare their pricing, BIN availability, API capabilities and onboarding requirements, then recommend the top three. ⸻ 2. Vendor Outreach & Follow-Ups Hermes should be able to: * Send emails to vendors * Request pricing and quotations * Ask predefined qualification questions * Read vendor replies * Extract important information * Maintain records of conversations * Automatically follow up if vendors do not respond * Notify the founder when important replies arrive * Prepare negotiation responses * Schedule sales/demo calls when appropriate Human approval can be required before sensitive or important outbound communications. ⸻ 3. Gmail & Google Workspace Integrate Hermes with: * Gmail * Google Calendar * Google Drive * Google Docs * Google Sheets Example workflows: * Read vendor replies * Add quotations to a spreadsheet * Create research documents * Schedule meetings * Maintain vendor comparison sheets * Save proposals and documents automatically ⸻ 4. Slack / Team Operations Hermes should be able to assist with internal team management. Examples: * Read relevant Slack messages * Identify outstanding tasks * Follow up with team members * Collect updates * Identify blockers * Send reminders * Generate daily summaries * Escalate unresolved issues to the founder Example: Every morning identify overdue tasks, ask responsible team members for an update, and send the founder a summary of blockers. ⸻ 5. Project & Task Management Build a simple system allowing Hermes to: * Track active projects * Track tasks * Track owners * Track deadlines * Remember project context * Follow up automatically * Detect overdue work * Produce daily/weekly project summaries This can use Google Sheets, Notion, Airtable, Basecamp, or another simple system if appropriate. ⸻ 6. Automation & Technical Work Hermes should be configured to: * Use terminal / shell * Run Python * Run JavaScript * Work with APIs * Read API documentation * Build scripts * Debug integrations * Create simple internal tools * Work with webhooks * Interact with n8n * Create or modify n8n workflows where possible Example: Research an API, build the integration, test it, identify errors and fix them without requiring the founder to guide every step. ⸻ 7. Browser Automation Set up reliable browser access so Hermes can: * Open websites * Navigate dashboards * Fill forms * Extract information * Download files * Perform repetitive browser tasks * Maintain authenticated sessions where technically possible The agent should attempt to solve browser-related problems independently before escalating. ⸻ 8. Multi-Agent / Model Routing We want different models used depending on the type of work. You should configure the system so that Hermes can use models such as: Primary agent * Claude Sonnet Complex reasoning / important decisions * Claude Opus * GPT reasoning models Coding * Codex or another strong coding model Cheap parallel research * Gemini Flash or another low-cost model The exact models can be adjusted based on availability and cost. We want intelligent model routing instead of sending every task to the most expensive model. ⸻ 9. Subagents Configure Hermes to delegate work to multiple agents when useful. Example: Main Hermes receives: Research the best attribution platform for our ecommerce business. It may delegate: Agent 1: Research pricing. Agent 2: Research Reddit feedback. Agent 3: Research integrations. Agent 4: Research API capabilities. Main Hermes should combine the results and produce the final recommendation. ⸻ 10. Memory / Company Knowledge Configure persistent memory so Hermes understands: * Our company * Our tools * Our preferred workflows * Previous vendor conversations * Previous research * Existing projects * Important decisions * Frequently used instructions The goal is to avoid repeatedly explaining the same context. ⸻ 11. Scheduled Jobs / Proactive Work Configure scheduled workflows such as: Morning operations report Check: * Outstanding projects * Overdue tasks * Important emails * Vendor responses * Team blockers Then send a concise summary. Vendor follow-ups Automatically follow up with vendors after a defined period. Weekly operations report Summarize: * Projects * Vendor research * unresolved issues * completed work * opportunities for automation ⸻ Safety / Approval System Not every action should happen automatically. Create an approval system. Hermes MAY perform automatically: * Research * Reading emails * Analysis * Drafting * Coding * Testing * Updating internal records * Creating reports * Checking project status * Routine follow-ups where approved Founder approval should be required for: * Sending sensitive external messages * Spending money * Entering contracts * Agreeing to pricing * Deleting important data * Changing production systems * Making major advertising/account changes * Sharing credentials * Financial transactions ⸻ Deliverables By the end of the project we expect: 1. Working Hermes installation Hermes installed and configured on our VPS/computer environment. 2. Model configuration Multiple AI models configured and tested. 3. Tool integrations At minimum, configure as many of the following as possible: * Browser * Gmail * Google Calendar * Google Drive * Google Sheets * Slack * n8n * Terminal * Python * APIs * Scheduled jobs 4. Hermes skills / workflows Create reusable workflows for: * Research * Vendor research * Vendor outreach * Vendor follow-up * Software comparison * Project tracking * Team follow-up * Technical troubleshooting * Automation development * Daily founder reporting 5. Testing Every major workflow must be tested with real example tasks. 6. Documentation Provide a short document explaining: * How the system works * How to start/stop Hermes * How to add models * How to add integrations * Where prompts/skills are stored * How to modify workflows * Known limitations 7. Handover Provide a short recorded screen-share/video explaining the final setup. ⸻ Timeline The project must be completed within: 3–4 days after starting. You should be available during this period to rapidly implement, test, fix problems and complete the system. Do not apply if you need several weeks to learn Hermes from scratch. ⸻ Budget Fixed budget: $100 USD This is intended as a fast MVP implementation using Hermes’s existing functionality, integrations, APIs, skills, prompts and automation capabilities. If the initial build is successful, there may be significantly more ongoing work improving the agent and building additional automations. ⸻ Who We Are Looking For You are a strong fit if you: * Have built AI agents before * Understand agentic workflows * Have worked with Hermes Agent or similar frameworks * Are comfortable working on Linux/VPS environments * Can work with APIs without extensive instructions * Understand OAuth and API authentication * Can debug technical issues independently * Know Python and/or JavaScript * Have experience with n8n / Make / Zapier * Understand LLM APIs * Understand prompt engineering * Can figure things out quickly * Can deliver working systems instead of only discussing solutions We care more about your ability to build and solve problems than formal qualifications. ⸻ When Applying Start your application with: HERMES Then answer the following: 1. Have you used Hermes Agent before? If yes, explain what you built. 2. What other AI agent frameworks have you used? 3. Which LLM APIs have you worked with? 4. Have you built MCP integrations? 5. Have you built n8n workflows? 6. Can you work with Gmail, Slack and Google APIs? 7. Can you work directly on a Linux VPS? 8. Give one example of an AI agent or automation system you independently built. 9. Which parts of the project above can you complete within 3–4 days? 10. Are you comfortable completing the MVP for the fixed $100 budget? 11. Send screenshots, GitHub repositories, Loom videos or examples of previous AI agent projects if available. Applications that simply say “I can do this” without explaining relevant technical experience will be ignored. Important We are not looking for someone to develop a brand-new AI agent framework from scratch. Hermes already provides much of the underlying infrastructure. Your job is to configure, integrate, automate, prompt, test and turn Hermes into a useful operational AI agent for our business as quickly as possible.