Job Description

At Proposales, we’re on a mission to shape the future of business proposals, replacing static documents by defining a new online standard fully powered by web technology (literally, a piece of internet infrastructure).

We’re looking for a Lead AI Engineer to help build intelligent capabilities into the Proposales platform — including proposal generation, structured content understanding, workflow automation, and developer tooling powered by modern language models.

This role sits at the intersection of language-model powered systems, product engineering, and platform infrastructure. You will turn modern AI capabilities into reliable production systems while helping shape how our engineering team builds with AI every day.

Quick overview

• Build production-grade language-model powered systems, agents, and workflow automation

• Design infrastructure for model orchestration and asynchronous workflows in a serverless, event-driven environment

• Work closely with Product and the Core platform team

• Stay hands-on with coding, experimentation, and system design

• Help define pragmatic practices for building and operating language-model powered features across engineering

Your mission:

• Design and build reliable AI-powered product capabilities using LLMs, retrieval systems, and intelligent workflows

• Write testable, maintainable, and scalable code with a strong focus on quality and long-term maintainability

• Continuously experiment, evaluate, and iterate on AI solutions to improve performance, accuracy, and user experience

• Develop reusable AI building blocks and internal tooling that accelerate product development

• Design and maintain AI infrastructure and workflows, including orchestration, background execution, and evaluation pipelines

• Help integrate AI deeply into both product features and internal developer workflows

• Contribute to shaping our engineering culture and workflows to make Proposales a place you look forward to in the morning

How you will work with AI at Proposales:

• Design and buildproduction-grade language-model capabilities, includingMCP-based integrations, standalone agents, and workflow-driven toolsthat solve real product problems. Establishservice layers and architectural patternsto make agentsreliable, observable, and reusablewithin aserverless, event-driven environment.

• Define howAI systems are tested and evaluated, improving reliability throughstructured evaluation, guardrails, and iterative experimentationfor non-deterministic systems.

• Improveinternal developer workflows and toolingusingAI-assisted development tools(e.g., Cursor, Claude) while maintaining highcode quality and maintainability.

• Work closely withinfrastructure concernssuch asorchestration, asynchronous execution, monitoring, and cost awareness, treating AI ascore platform infrastructurerather than standalone features.

• Applystrong architectural thinkingto language-model systems and help definepractical patterns for model-driven infrastructurewithin aserverless ecosystem.

What we believe in:

We use a serverless, continuous-delivery infrastructure hosted on Vercel, built using Next.js and written in TypeScript.

We treat AI as an engineering discipline — reliability, observability, evaluation, and thoughtful user experience matter as much as model capability.

We stay pragmatic when working with AI — we explore and adopt external vendors and tools when they provide clear value, while designing our systems so we can evolve and adapt as the ecosystem changes.

AI at Proposales is treated as infrastructure as much as product — including workflows, orchestration, evaluation and automations.

When deciding on a tech stack for a new sub-project, we usually take the opportunity to assess new tech early and plan for long-term benefits and maintainability over short-term goals. But we also value simple solutions over smart ones, and embrace test-driven development and pair programming, favoring higher code quality over faster development times.

We prefer monorepos with multiple smaller projects, instead of one monolithic application. We avoid code duplication by extracting core logic and UI into separate internal libraries.

We use AI where it clearly helps — in product and internal tooling — while keeping automation and human judgment in the right balance.

Core AI Skills

• Modern language models (LLMs/SLMs) and transformer fundamentals

• Retrieval-Augmented Generation (RAG) and retrieval systems

• Embeddings, vector databases, and semantic search

• Agent architectures and autonomous AI systems

• Prompting strategies and structured prompting techniques

• AI evaluation, experimentation, and model quality assessment

• AI infrastructure and production AI system design

• Workflow orchestration and background job systems

• Model Context Protocol (MCP) and modern AI integration patterns

Technologies You’ll Work With

• Languages: Python, TypeScript / JavaScript

• Backend: Node.js, Next.js

• Data: PostgreSQL, vector databases, embedding models

• AI Tooling: Vercel AI SDK and related AI development frameworks

• Infrastructure: Serverless architectures and cloud-native deployments (Vercel)

• Systems: Event-driven architectures and workflow orchestration platforms

A Bit About You

• You are a strong engineer who enjoys building real products, not just prototypes

• You are pragmatic about AI, balancing experimentation with reliability and real user value

• You enjoy working closely with product and design to shape ideas into practical solutions

• You take ownership and are comfortable leading technical decisions

• You enjoy mentoring engineers and helping teams adopt new technologies

• You value simple, maintainable systems over unnecessary complexity

• You care about code quality through reviews, testing, and collaborative development

• You communicate clearly and work well in cross-functional teams

• You are curious and enjoy learning new technologies in the rapidly evolving AI space

• Strong English communication skills; Stockholm-based and comfortable in an office-first environment.