Senior AI / Machine Learning Engineer (m/f/d)

pplwise GmbH; Berlin; Remote; EUR 110,000 - 150,000; Full time

Benefits offered by this job

Remote work

Team Days in Mannheim

Job summary

pplwise GmbH is seeking a Senior AI/ML Engineer to own the models powering our platform that transforms drawings, datasheets, and catalogue data into structured part data for intelligent purchasing decisions.

You will collaborate with the engineering team, fine-tune models, define datasets, and build evaluation gates to ensure reliable model deployments. The role is remote with monthly in-person Team Days at our Mannheim office.

Qualifications

• Experience building evaluation protocols for ML models.
• Ability to define datasets and ground truth for model training.
• Knowledge of training and serving pipelines for ML models.

Responsibilities

• Own models that decide what the platform tells a customer and the evidence behind it.
• Train and fine-tune models on technical documents and related data.
• Develop evaluation gates and automatic checks for model changes.
• Ensure parity between training and production features and preprocessing.
• Stand up multi-tenant model deployments with per-customer thresholds.
• Communicate decisions and progress clearly with stakeholders.

Skills

ML engineering

Model evaluation

Data preprocessing

Open-weight models

Tools

PyTorch

Transformers

Job description

AI Agents to Bridge the Gap between Engineering & Purchasing to reduce manufacturing waste

My co-founder Ian and I combine over 15 years of industry experience, working with companies like SEW-Eurodrive, Terex, Bosch and Carl Zeiss. We co-founded Resourcly, a Project A-backed, industrial tech start-up to help industrial manufacturers streamline complex portfolios to improve EBIT while preventing and reducing idle inventory from being scrapped.

With our solution, we want to solve industry problems that we have tried to address for the last decade in the manufacturing companies we worked for, but that have so far been out of reach. With new technology and AI, that can finally happen.

Complex processes and products, lack of engineering talent and massive labour costs as well as a lack of intelligent automation have remained the industry's most pressing challenge. With LLMs and agents, we finally have the tools to solve this.

Our vision is that future purchasing and engineering processes will be orchestrated and run by an AI workforce - ensuring reliable and cost-efficient procurement and flow of goods in an uncertain and complicated world with finite resources.

We're already live with global manufacturing leaders in Europe

Backed by one of Europe's top early-stage funds (Project A)

Growing fast and hiring exceptional talent to join early

Our goal is to build a small, but excellent team, therefore we are looking for a Senior AI/ML Engineer. If you are excited to work in a very early-stage environment alongside the founding team, stellar data scientists and engineers, using AI at the product core to solve meaningful problems - please reach out.

The Role

As our Senior AI/ML Engineer, you will own the models that decide what our platform tells a customer, and the evidence that they are right. We turn technical drawings, datasheets and messy catalogue data from some of Europe's largest manufacturers into structured, comparable part data, and then answer the question everything rests on: are these two parts the same part? Customers make purchasing and engineering decisions on those answers, so a wrong answer costs more than a slow page.

This is a second senior hire into an existing applied-ML practice, with models already serving customers. You will work side by side with the engineer who owns our model pipelines today, report to our Head of Engineering, and collaborate closely with backend engineers, product and our customer-facing team. You will build and fine-tune models, and you will build the thing we are still missing: an evaluation strong enough to decide whether a model ships. This is neither a research role nor a platform role. You define the problem from the data, build the dataset, train the model, prove it, and put it on a GPU.

• Build evaluation that can actually decide: Design leakage-safe splits, per-class quotas for the rare classes that matter, and power calculations before comparisons, then turn it into an automated gate and a durable record so every model change is adjudicable
• Fine-tune what we can serve: Train open-weight vision-language models on technical documents and own the cheaper models in front of them, select on the task metric rather than the loss curve, and kill your own model when a simpler approach wins
• Own datasets and ground truth: Define label taxonomies precisely, build ground truth where none exists, repair labels with clean provenance, and close the loop so human corrections become training data
• Keep training and serving identical: Own the parity between how features and preprocessing are computed at training time and in production, as executable checks rather than conventions
• Own serving, cost and throughput: Take a training run to a reproducible serving artefact on GPU, measure like for like, and pick the cost lever that actually applies
• Roll the pipeline out to the next customer: Stand up models for new customers in a multi-tenant setup, with thresholds and model identity that travel per customer and tenant isolation as a hard requirement
• Connect the model to the customer: Understand which mistakes are cheap and which are expensive, review real output with customers and colleagues, and explain a model's limits honestly to the people acting on it
• Intellectually honest: You report the number that makes you or your model look worse, and say plainly when a result cannot answer the question being asked
• Take full ownership: From dataset and training through serving, monitoring and rollback in production
• Bring startup mindset with high agency: Comfortable with ambiguity, because defining the problem, the benchmark and the metric is the first job
• Data before model: You look at the labels, the data and the preprocessing before reaching for a bigger model
• Low ego, strong teammate: You can inherit somebody's else code, name its strengths as readily as its faults, and improve it in place rather than rewriting it
• Curious about the domain: You want to know what a specification is and what the customer does with your answer
• Proactive communication: You write decisions down with their reason and their measurement, and share progress, blockers and trade-offs early
• Start: As soon as possible
• Compensation: Top of market (We're constrained by bandwidth, not demand - great time to negotiate)
• Team Days: The position is remote, with an expectation to attend monthly in-person Team Days at our Mannheim office.