Job Description

About Welvaart

At Welvaart, we create technology solutions that put people at the center.
Our close leadership style and flexible culture of growth empower our teams and elevate the quality of our delivery. We combine rigor, innovation, and empathy to drive projects that transform businesses and build lasting relationships of trust.

We complement this vision with a performance‑driven Digital Marketing offering, helping companies strengthen their visibility, enhance their online presence, and accelerate growth through smart, measurable strategies.

Project

You will build intelligent features such as classification, extraction, summarization, and action orchestration powered by large language models. You will design embedding and retrieval pipelines (RAG, semantic search), create robust data pipelines for training and evaluation, and define clear evaluation metrics and quality gates to ensure reliable LLM behavior in production.

You will work hands-on with inference runtimes such as ONNX Runtime and TensorFlow Lite, benchmarking performance across CPU, GPU, NPU, and DSP environments, and optimizing deployments for latency, cost, and reliability—including in constrained or embedded systems. Collaborating with engineering, data, and MLOps teams, you will integrate models into real-world APIs and production systems while continuously experimenting with prompts, architectures, and model choices.

Role

• Build and integrate LLM-powered features (classification, extraction, summarization, actions).
• Integrate models with inference runtimes (such as ONNX Runtime, TensorFlow Lite / LiteRT).
• Benchmark and validate model performance across different hardware backends (CPU, GPU, NPU, DSP).
• Design embedding and retrieval pipelines (RAG, semantic search).
• Create and maintain data pipelines for training and evaluation.
• Define evaluation metrics and quality gates for LLM behavior.
• Optimize inference for latency, cost, and reliability.
• Integrate models into production systems and APIs.
• Run experiments to evaluate prompts, models, and architectures.

We are looking for

• Strong experience with LLMs and NLP systems
• Hands-on experience with embeddings and vector databases
• Strong Python skills and ML frameworks
• Experience building production data pipelines
• Solid understanding of evaluation and regression detection
• Experience with RAG architectures
• MLOps or monitoring experience
• Experience with model calibration and accuracy/latency trade-off analysis.
• Hands on experience deploying models on edge or embedded devices (constrained environments)

What you can discover with us?

• Be part of a tech start-up
• Different scopes of project in different sectors
• Structure of fairness and equity salary (Consultant Profile)
• Training & Certification
• Career Path management
• More than 30 Partnerships

UNLEASH THE POWER OF YOUR CAREER