Full-Stack Software Engineer (AI & Agentic Systems)

Full-Stack Software Engineer (AI & Agentic Systems)

Job title Full-Stack Software Engineer (AI & Agentic Systems)

Location Ukraine (Remote)

Salary Competitive, depending on experience

Reports to Anstar Head of IT

Tech Stack

• Frontend & Backend: TypeScript, React, TanStack Start, Node.js / Bun, PostgreSQL
• AI & Agentic Tooling: Agentic Coding, Skills, MCPs, AI Gateways, Hermes, Codex (open to benchmarks and experiments with alternative tools)
• Automation & Enterprise: Dynamics 365 Sales Pro, Dataverse, Business Central, Power Apps, N8N, Custom LLM Runtimes

About the Company

Anstar is a rapidly growing speciality chemicals distributor with operations across the UK, Europe and the US. With deep cultural and commercial expertise, we serve as a trusted bridge between Eastern manufacturers and global markets. We have quickly established ourselves as leaders in the UV-curing sector and have also expanded into Nutraceuticals. Our agile, lean structure prioritises transparency, openness, and a customer-first approach, driven by ambitious, high-potential individuals.

Highlights

• The client is optimizing and transforming internal systems and operational processes through automation and modern AI tooling.
• Direct involvement with both custom-built software and enterprise platforms to streamline daily business operations.
• Practical engineering culture focused on ownership, sensible experiments, and measurable results.

The Opportunity

We are looking for a hands-on Full-Stack AI Engineer to develop and maintain our internal product ecosystem and automated workflows. This role is built for an engineer who can self-organize, operate independently, and take full ownership of complex system modules.

You will build features end-to-end, implement agentic workflows, and connect our systems through APIs, custom tools, and automated pipelines. The environment spans enterprise platforms (Dynamics 365, Business Central, Power Apps) alongside custom internal tools built with a modern stack including TanStack Start, Bun/Node.js, and custom LLM runtimes.

We actively work with Hermes and Codex for agentic workflows and coding automation, and we welcome grounded experimentation with other tools whenever backed by clear performance benefits.

Key Responsibilities

• Develop, scale, and maintain complex internal systems using our modern web and database stack.
• Implement and refine agentic features, custom skills, and MCP integrations to automate operational flows.
• Connect custom applications with internal enterprise platforms and third-party services via clean API architectures and N8N workflows.
• Take ownership of discrete modules: investigate edge cases, handle latency/reliability issues, and deliver functional increments end-to-end.
• Evaluate and benchmark agentic coding workflows (primarily Codex and Hermes) to boost engineering velocity and system quality.
• Collaborate directly with stakeholders to clarify requirements, define practical scopes, and ship usable tooling.

Requirements

• Strong production experience with TypeScript, React, and relational databases (PostgreSQL).
• Solid foundation in modern full-stack application architecture (SSR/hybrid frameworks, runtime environments).
• Hands-on experience with agentic software engineering concepts:
• Systematic prompt engineering and structured system messages
• Developing and configuring Agent Skills
• Building and integrating MCPs
• Routing and resilience with AI Gateways
• Practical usage of Codex and Hermes (or demonstrated proficiency with equivalent agentic frameworks)
• Proven ability to take ownership of complex technical tasks and see them through to delivery.
• Clear communication skills with the ability to discuss technical decisions and trade-offs.

Desirable / Nice to Have

• Experience with .NET / C#.
• Familiarity with Microsoft Business Applications (Business Central, Dynamics 365 Sales Pro, Dataverse, Power Apps).
• Experience building or maintaining N8N automation pipelines.
• Practical ML / LLM engineering background:
• Self-hosting and running local models
• Model fine-tuning (LoRA / QLoRA)
• Experience with PyTorch