Job VC
Architect of AI-agent systems
Technologies
Description
Our Client is an entrepreneur with an operational business in the field of B2B trade and logistics and foreign economic activity, which is growing in volume. Instead of scaling the team by hiring, I am building automation based on AI agents — so that it takes over operational processes with minimal human intervention.
Some of the agents have already been launched independently. They work about 80% — but I need a level of
97-100%
with professional architecture, harness, evals and everything that makes such systems stable in production.
We are looking for a tech lead / architect who fully understands the architecture of AI-agent systems and is able to bring them to a production-level of reliability. Your “development team” is mainly AI coding agents; your task is to design, set tasks, control quality and independently bring the solution to combat operation.
This is not a one-time project. This is a full-time, long-term role with the potential to expand into the product direction if appropriate understanding is developed in the process.
Responsibilities
First
3-6 months:
operational automation of the business to
97-100%
reliability
Audit of current processes and formation of an implementation roadmap
Design of the architecture of the agent system: harness, multi-agent workflows, tool use, RAG, evals, prompt and version control
Bringing already launched agents to the production level
Gradual coverage of processes:
Invoice processing — preparation, entry into the system
Logistics — preparation and verification of documents, status updates
CRM — fully automated provision of data relevance and correctness
Commercial offers — calculations and CP on templates
Payment control — sync with banks, payment distribution, payment discipline control
Analytics — cashflow, unit-economics, margin, capital use
Communication with customers / suppliers — at the start in passive mode to accumulate context on processes
Sourcing of suppliers — in the final turn, as frontline process with increased attention
Requirements
AI / LLM agent systems — must-have #1.
Deep expertise: LangGraph (or equivalent), RAG, evals, tool use, multi-agent workflows, harness, prompt and version control. Real experience in bringing agent systems to production with measurable reliability.
Mindset of a business automator.
You see the process — and immediately understand how to turn it into an agent workflow.
AI-first product thinking.
You think in terms of systems that scale, not one-time scripts for each task.
Software architecture / backend.
High-quality design and architecture discipline.
Local inference — will be a plus.
Experience in deploying local models, optimizing for hardware, building hybrid local + API pipelines.
Some of the agents have already been launched independently. They work about 80% — but I need a level of
97-100%
with professional architecture, harness, evals and everything that makes such systems stable in production.
We are looking for a tech lead / architect who fully understands the architecture of AI-agent systems and is able to bring them to a production-level of reliability. Your “development team” is mainly AI coding agents; your task is to design, set tasks, control quality and independently bring the solution to combat operation.
This is not a one-time project. This is a full-time, long-term role with the potential to expand into the product direction if appropriate understanding is developed in the process.
Responsibilities
First
3-6 months:
operational automation of the business to
97-100%
reliability
Audit of current processes and formation of an implementation roadmap
Design of the architecture of the agent system: harness, multi-agent workflows, tool use, RAG, evals, prompt and version control
Bringing already launched agents to the production level
Gradual coverage of processes:
Invoice processing — preparation, entry into the system
Logistics — preparation and verification of documents, status updates
CRM — fully automated provision of data relevance and correctness
Commercial offers — calculations and CP on templates
Payment control — sync with banks, payment distribution, payment discipline control
Analytics — cashflow, unit-economics, margin, capital use
Communication with customers / suppliers — at the start in passive mode to accumulate context on processes
Sourcing of suppliers — in the final turn, as frontline process with increased attention
Requirements
AI / LLM agent systems — must-have #1.
Deep expertise: LangGraph (or equivalent), RAG, evals, tool use, multi-agent workflows, harness, prompt and version control. Real experience in bringing agent systems to production with measurable reliability.
Mindset of a business automator.
You see the process — and immediately understand how to turn it into an agent workflow.
AI-first product thinking.
You think in terms of systems that scale, not one-time scripts for each task.
Software architecture / backend.
High-quality design and architecture discipline.
Local inference — will be a plus.
Experience in deploying local models, optimizing for hardware, building hybrid local + API pipelines.