pMost AI products are wrappers. We're building the real thing, an agent that takes on genuine tasks for everyday users: running errands, managing workflows, holding context across long and complex conversations. Reliable by design, not by luck. /ppWe're small, we move fast, and the ML layer is the product. We need someone to own it. /ppstrongThe role /strong /ppYou'll bridge research and production, taking ideas and turning them into systems that run at scale, stay reliable, and get better over time. Full-stack ML ownership: from raw data to deployed model. /ppDay to day that looks like: /pulliBuilding end-to-end pipelines across data, training, evaluation, and inference /liliAdapting and fine-tuning models with modern techniques: LoRA, QLoRA, SFT, DPO, distillation /liliArchitecting inference systems that hold up under real latency and cost constraints /liliCreating data pipelines that produce high-quality synthetic and real-world training data /liliRunning evaluation that goes beyond benchmarks: robustness, safety, bias, production behaviour /liliOwning deployment: GPU optimisation, quantisation, memory efficiency, scaling /liliWorking directly with application engineers so ML integrates cleanly into backend, mobile,



and desktop /li /ulpstrongYour skills and experience /strong /pulliDeep understanding of deep learning and transformer architectures /liliProven experience training, fine-tuning, or shipping large-scale models in production /liliStrong with at least one major ML framework (PyTorch, JAX) and quick to pick up others /liliFamiliar with distributed training and inference tooling: DeepSpeed, FSDP, Megatron, ZeRO, Ray /liliEngineering discipline: code that's readable, robust, and maintainable /liliExperience optimising for GPU constraints: quantisation, mixed precision, memory /liliComfortable taking ownership of ambiguous problems from zero to one /liliShips, iterates, learns from production /li /ulpstrongNice to have /strong /pulliLLM inference frameworks: vLLM, TensorRT-LLM, FasterTransformer /liliRLHF: PPO, DPO, ORPO /liliOpen-source contributions to ML or systems libraries /liliScientific computing, compiler, or GPU kernel experience /li /ulpAt a big company, ML work gets absorbed into a machine. Here, your systems are the product. You'll work closely with research and engineering leadership, have real influence over how the architecture evolves, and see the direct impact of your work on users. If you want to build ML infrastructure that actually matters, this is it. /p #J-18808-Ljbffr

📌 Senior Machine Learning Engineer (España)
🏢 Digital Waffle
📍 España