GenAI Tech Lead

Hexecapital; Wrocław; Hybrid; PLN 264,000 - 316,000; Full time

Benefits offered by this job

Modern, pet-friendly office

Central Wrocław location

Autonomy and fast decision-making

Competitive B2B pay

Job summary

Hexecapital is building an AI platform in Poland to deliver GenAI solutions for large clients. You will lead architectural decisions, select models, and implement working AI agents with a focus on latency, quality, and cost.

You’ll prototype approaches, maintain high code quality, and collaborate closely with product owners and clients to push platform capabilities forward.

Qualifications

• Strong commercial experience building production-grade solutions based on LLMs.
• Hands-on experience with RAG, context engineering, tool/function calling, and LLM evaluation.
• Ability to analyze model quality, latency, and cost for technical decisions.
• Solid software engineering background with architecture-level thinking.

Responsibilities

• Developing our platform for building AI Agents.
• Designing architecture of production-grade AI solutions.
• Selecting models for use cases and trade-offs between quality, speed, cost, and stability.
• Designing and improving evaluation frameworks for models and GenAI solutions.
• Diagnosing latency, hallucinations, context, and reliability issues.
• Solving quality and performance problems in production AI systems.
• Making architectural decisions around RAG, context engineering, tool calling, and outputs.
• Designing new platform features and assessing impact on existing solutions.
• Prototyping and comparing approaches before production.

Skills

LLMs

GenAI

RAG

Context engineering

Tool calling

Latency optimization

Software engineering

PoC implementation

Tools

Node.js

TypeScript

Vue

RAG-based solutions

Job description

Using our own AI platform, we build GenAI solutions for some of the largest companies in Poland, includingTVN, Żabka, Biedronka, Santander Consumer Bank, and many more .

We’re looking for someone for whom LLMs and GenAI are a genuine area of expertise, backed by hands-on experience with production-grade solutions. Someone who understands both their capabilities and limitations and can turn that knowledge into sound technical decisions.

What this role looks like:

• What architecture makes the most sense for a specific use case?
• Where does voicebot latency come from, and how do you reduce it?
• How do you compare models in a meaningful way?
• When is a prompt change enough, when do you need RAG, and when is the real problem deeper in the architecture?

These are the kinds of questions you’ll work on every day. You’ll choose the right direction, turn it into a working solution, and lead its implementation.

What you’ll be responsible for:

• developing our platform for building AI Agents,
• designing the architecture of production-grade AI solutions,
• selecting models for specific use cases and making informed trade-offs between quality, speed, cost, and stability,
• designing and improving evaluation frameworks for both models and complete GenAI solutions,
• diagnosing issues related to latency, response quality, hallucinations, context, and reliability,
• solving quality and performance problems in production AI systems,
• making architectural decisions around RAG, context engineering, tool/function calling, structured outputs, and other LLM-related components,
• designing new platform features and assessing their impact on existing solutions,
• prototyping and comparing different approaches before deciding what should go into production,
• maintaining implementation quality through code reviews and hands-on support with the most technically challenging parts,
• working closely with our technical Product Owner on platform development and new experiments.

You’ll be a great fit if:

• you have strong commercial experience building production-grade solutions based on LLMs,
• you understand the strengths and limitations of different models and can choose the right one for a specific problem,
• you have hands-on experience with RAG, context engineering, tool/function calling, and LLM evaluation,
• you can analyse model quality, latency, and cost, and use those insights to make technical decisions,
• you have a solid software engineering background,
• you can think about systems as a whole and make architectural decisions with their long-term consequences in mind,
• you can translate business problems into technical solutions and clearly communicate both their capabilities and limitations,
• you’re comfortable working directly with code and can independently validate a hypothesis through a PoC or implementation,
• you thrive in a high-autonomy environment and enjoy solving problems that don’t come with a ready-made playbook,
• *experience with Voice AI will be a big plus, especially voicebots, STT/TTS, streaming, real-time processing, or latency optimisation.

Our stack:

• the core of our platform is built in Node.js with TypeScript,
• the frontend is built in Vue,
• in the AI layer, we work with different LLMs and RAG-based solutions,
• we’re not tied to a single model or provider – we test, compare, and choose the solution that best fits the problem.

What we offer:

• the opportunity to co-create a proprietary AI platform built for large-scale enterprise implementations,
• hands-on work with LLMs, GenAI, and technologies that are rapidly changing how digital products are built,
• real influence over the platform architecture and the direction of our AI solutions,
• freedom to experiment with, test, and implement new technologies without unnecessary bureaucracy,
• a high level of autonomy, trust, and space to make your own technical decisions,
• a team of people who enjoy solving difficult technical problems and aren’t afraid to challenge established approaches,
• fast decision-making and direct collaboration with the CEO, COO, and people responsible for product and clients,
• collaboration with clients implementing AI at scale and bringing genuinely ambitious technical challenges,
• a culture based on partnership, transparency, and the freedom to challenge the status quo,
• a modern, pet-friendly office in central Wrocław,
• compensation of up to 26 000 PLN net/ month on B2B, depending on experience.

If you want to influence not only how something is built, but first and foremost what should be built and why, we’d love to meet you!

Just tell us who you are and what you’re great at.