Senior AI Engineer for a Deep Agents Platform and Scalable SaaS
Budget: €30 – €250 EUR
Hello,
We are looking for a senior AI/full-stack engineer, or a small senior team, to develop a proprietary AI agent platform built with Deep Agents, LangChain and LangGraph.
Our long-term goal is to create a scalable B2B SaaS with one of the most advanced enterprise agent systems in the market: Super Agents capable of completing long-running and complex assignments, coordinating specialized subagents, maintaining persistent memory and context, connecting to APIs and MCP servers, operating browsers, recovering from failures, and working under configurable levels of autonomy and human approval.
Despite the technical sophistication, the product must be simple for the end customer. It will include a complete administrative dashboard where non-technical users can configure and manage agents, skills, models, knowledge, memory, integrations, MCP connections, tasks, approvals, costs, permissions and recurring automations without changing code during normal operations.
The reference architecture includes:
* Python, Deep Agents, LangChain and LangGraph;
* FastAPI and asynchronous workers;
* PostgreSQL, pgvector, checkpoints and persistent memory;
* MCP, external APIs and webhooks;
* Playwright and/or Browser Use;
* Docker Compose and deployments on VPS or AWS;
* Next.js, TypeScript and a complete administration dashboard;
* authentication, organizations, RBAC and audit logs;
* observability, evaluations, cost tracking and security;
* support for multiple LLM providers and models.
We already have a comprehensive technical and architectural specification. We do not intend to build the entire SaaS immediately. The project will be contracted through independent phases:
1. Technical audit and validation of the existing architecture.
2. Proof of concept for the core Super Agent.
3. Private MVP for use within our own businesses, including the administration dashboard.
4. Reproducible deployment package for VPS or AWS without creating separate code forks for each customer.
5. Internal validation, recovery testing, security testing and cost analysis.
6. Deployment for 2 to 5 pilot customers in dedicated and isolated environments.
7. Transformation of the validated product into a scalable multi-tenant SaaS.
8. White-label capabilities, partner management and a possible agent marketplace.
We require proven experience with AI agent systems, LangChain/LangGraph, Python, durable workflows, persistent memory, tool calling, MCP, browser automation, PostgreSQL, Docker, security and SaaS architecture. Experience with Next.js/TypeScript, or the ability to collaborate with a senior frontend engineer, is also important.
The engagement must meet the following conditions:
* all code must be stored in our company GitHub from the beginning;
* technical documentation and automated tests are mandatory;
* the architecture must remain modular and independent from any single LLM or cloud provider;
* versions and dependencies must be locked and documented;
* all original code, dashboards, workflows, skills, integrations and documentation specifically developed for the project must be assigned to our company;
* third-party open-source components must retain and comply with their respective licences;
* we must be able to stop or change suppliers at the end of each phase;
* all credentials, infrastructure, documentation and technical knowledge must be handed over.
At this stage, we are requesting a fixed quotation for Phases 1 and 2, together with an indicative estimate for the remaining phases.
Please include:
* relevant experience and similar projects;
* portfolio, GitHub profile or case studies;
* proposed team and responsibilities;
* initial technical approach;
* timeline and pricing separated by phase;
* availability;
* support and maintenance terms.
We are not looking for a basic LLM integration or a simple chat interface. We are looking for a senior professional capable of helping us build an autonomous, durable, auditable and commercially scalable enterprise agent platform.
We can provide the complete technical specification after an initial review of your profile.
We are looking for a senior AI/full-stack engineer, or a small senior team, to develop a proprietary AI agent platform built with Deep Agents, LangChain and LangGraph.
Our long-term goal is to create a scalable B2B SaaS with one of the most advanced enterprise agent systems in the market: Super Agents capable of completing long-running and complex assignments, coordinating specialized subagents, maintaining persistent memory and context, connecting to APIs and MCP servers, operating browsers, recovering from failures, and working under configurable levels of autonomy and human approval.
Despite the technical sophistication, the product must be simple for the end customer. It will include a complete administrative dashboard where non-technical users can configure and manage agents, skills, models, knowledge, memory, integrations, MCP connections, tasks, approvals, costs, permissions and recurring automations without changing code during normal operations.
The reference architecture includes:
* Python, Deep Agents, LangChain and LangGraph;
* FastAPI and asynchronous workers;
* PostgreSQL, pgvector, checkpoints and persistent memory;
* MCP, external APIs and webhooks;
* Playwright and/or Browser Use;
* Docker Compose and deployments on VPS or AWS;
* Next.js, TypeScript and a complete administration dashboard;
* authentication, organizations, RBAC and audit logs;
* observability, evaluations, cost tracking and security;
* support for multiple LLM providers and models.
We already have a comprehensive technical and architectural specification. We do not intend to build the entire SaaS immediately. The project will be contracted through independent phases:
1. Technical audit and validation of the existing architecture.
2. Proof of concept for the core Super Agent.
3. Private MVP for use within our own businesses, including the administration dashboard.
4. Reproducible deployment package for VPS or AWS without creating separate code forks for each customer.
5. Internal validation, recovery testing, security testing and cost analysis.
6. Deployment for 2 to 5 pilot customers in dedicated and isolated environments.
7. Transformation of the validated product into a scalable multi-tenant SaaS.
8. White-label capabilities, partner management and a possible agent marketplace.
We require proven experience with AI agent systems, LangChain/LangGraph, Python, durable workflows, persistent memory, tool calling, MCP, browser automation, PostgreSQL, Docker, security and SaaS architecture. Experience with Next.js/TypeScript, or the ability to collaborate with a senior frontend engineer, is also important.
The engagement must meet the following conditions:
* all code must be stored in our company GitHub from the beginning;
* technical documentation and automated tests are mandatory;
* the architecture must remain modular and independent from any single LLM or cloud provider;
* versions and dependencies must be locked and documented;
* all original code, dashboards, workflows, skills, integrations and documentation specifically developed for the project must be assigned to our company;
* third-party open-source components must retain and comply with their respective licences;
* we must be able to stop or change suppliers at the end of each phase;
* all credentials, infrastructure, documentation and technical knowledge must be handed over.
At this stage, we are requesting a fixed quotation for Phases 1 and 2, together with an indicative estimate for the remaining phases.
Please include:
* relevant experience and similar projects;
* portfolio, GitHub profile or case studies;
* proposed team and responsibilities;
* initial technical approach;
* timeline and pricing separated by phase;
* availability;
* support and maintenance terms.
We are not looking for a basic LLM integration or a simple chat interface. We are looking for a senior professional capable of helping us build an autonomous, durable, auditable and commercially scalable enterprise agent platform.
We can provide the complete technical specification after an initial review of your profile.