AI/ML Jobs

The artificial intelligence, machine learning, and deep learning jobs

NLP Data Scientist and Data Engineer at Meeshkan (Helsinki, Finland)

We are Meeshkan, winners of the 2018 Slush 100 pitch competition and recently voted #1 on Wired's list of Finland's best startups ( We got there because we're building a fake version of the internet, and we are using NLP to do it. We are looking for a talented and passionate NLP engineer comfortable working with semi-structured data (JSON, XML, markdown, etc) to curate open datasets and lead the development of our next generation of world-class NLP algorithms.

Is this Really a Data Science Job?

Yes. Some companies advertise data science jobs but are really looking for someone to manage their Excel spreadsheets. We are a PyTorch house that has managed deployments of 500-computer clusters doing scraping and data processing. We have four in-house GPUs that keep us warm during the winter. We believe in solid data pipelines, organized experimentation, and pushing to production as fast as possible.

But Seriously, Will I Be Doing Machine Learning?

Yes. Not only that, but you will be leading our efforts in Machine Learning. Which means that in addition to adapting and designing the models, you will be building datasets, finding effective ways to sanitize them, and working with our partners like nVIDIA on how to apply state-of-the-art algorithms to the unique problem we are solving.

So What Problem Are You Solving?

Meeshkan is a professional service company that helps other companies make fake versions of their APIs and Microservices. Why fake? Because no one wants to do testing and experimentation on production APIs. Companies the world over use reliable simulations of their infrastructure in order to create new business and bring products to market faster. And they depend on us, and you, to make that happen.

Making reliable fake APIs means getting lots of things right. One of them is building adaptative generative models based on NLP technologies, like random forest, Bayesian inference, and RNNs. It also means knowing where the boundaries of ML lie and being comfortable writing glue code, like heuristics and plain-old-functions, to make and end-to-end solution work. It also requires packaging models in containers and sharing them with developers inside and outside Meeshkan.

Your Role in Meeshkan

Meeshkan is rebooting its data science division. That means first and foremost understanding what we've done (don't worry, we will onboard you), using existing processes and infra that you deem appropriate, and otherwise building new datasets, pipelines, models and deployment channels from the ground up.

In doing so, we also value community contributors that publish their work in academic journals and present it at conferences. If you are a postdoc or interested in one, this role could be particularly relevant for you.

Company Culture

Your day-to-day work would be fairly independent, but the entire team is always available for collaboration as a sounding board. We are low on formalities and high on substance, and in addition to our operational team, we have a broad network of advisors and investors that are always willing to help out in ideation and introductions. We have a generous office hours and vacation policy, but we are one of the most high-intensity and high-output companies in town due to the fast pace of our industry. Because of this, we are also hyper-sensitive to stress and wellbeing, and offer employees ample opportunities to recharge and feel good.


We are committed to fostering a diverse workplace, and we make it a special point to welcome candidates from underrepresented groups in tech companies. We know that, if you are part of one of these groups, you may have suffered from systemic industry-wide biases that make it more difficult for you to shine than other folks. We don't like that, and we are committed to evaluating every candidate on their own merits given their unique experiences.

Parting Shot

We are a data science company that understands the power and limits of machine learning. It is linked directly to our value prop and bottom line, and we put you and your work front and center in achieving those goals.