Data Scientist, LLMs and Prompt Engineering

NOTICE: ONLINE RECRUITMENT PROCESS

LOCATION: KRAKÓW OR REMOTELY FROM POLAND

SALARY FOR A MID: 17 500 – 30 000 PLN gross/monthly

SALARY FOR A SENIOR: 26 000 – 35 000 PLN gross/monthly

Every month we are proud to be home to 300 million users around the world. Brainly’s knowledge base consists of hundreds of millions of Q&A content in more than 12 languages and covers a broad spectrum of educational subjects at different grades.

Our AI strategy and roadmap are investing more and more in our capacity to best exploit modern LLMs and build domain-specific layers around them.

Being able to design and craft optimal prompts is fundamental for our success.

We would like to establish a Center of Excellence (CoE) for prompt engineering which is part of the broader AI Research team but with a specific focus on being up-to-date with the latest advancements in this field and acting as a go-to place for getting prompts tailored for each single application at Brainly, both for projects within and outside AI Services department.

The CoE will dedicate efforts to constantly researching and experimenting with the S-O-T-A and applying it to our business domain.

Thus, supporting production delivery teams to remain focused on meeting their deadlines and developing robust and high-quality ML production systems, while the CoE of Prompt Engineering figures out and optimizes how to leverage available LLMs for their delivery goals.

By centralizing this ownership we can build best-in-class expertise serving the whole company while the rest of the ML practitioners can focus on solving other ML problems such as classification of images and text, object detection, document understanding, recommender systems, and more.

ROLE OVERVIEW

As a Data Scientist in Prompt Engineering you will have the chance to work with top-class scientists, engineers, and domain experts, and to drive the data science and research processes of our LLM-based product features end-to-end.

The ideal candidate is an enthusiast of the educational domain with a blend of coding, machine learning, and statistics skillset, and most importantly motivated to work full-time to master the art of prompt engineering.

WHAT YOU’LL DO

• Research and experiment with the latest state-of-the-art methods applied to Brainly data and education domain
• Train and share knowledge with the rest of ml practitioners on those advances
• Implement novel techniques of language understanding, knowledge representation, and content generation
• Develop a catalog of default prompts, and models parameters, ready-to-use for a variety of applications
• E.g. question answering, personalization, classification, tagging, entity extraction, summarization, paraphrasing, text cleaning, ranking/comparison
• Rapidly provide new prompts, or optimize existing ones, for specific functionalities of Brainly’s product
• e.g. Ginny, AI tutor, and others
• Provide suggestions for new utilization of LLMs as part of Brainly product features or optimization of internal processes
• Provide consulting, initial research, and recommendations of which prompts to try as part of the R&D of planned production ML systems
• Identify strengths and weaknesses of LLMs (off the shelf and proprietary), content characteristics, or users’ intentions, and report insights related to opportunities for improvement
• Thus informing AI strategy and roadmap
• Develop methodologies and procedures for the evaluation of LLM-generated content
• Partner and train the AI Operations team, and other Brainly teams, on quality assurance related to LLM-generated content
• Enable the generation of high-quality labeled data
• Enable Brainly employees who are dealing with LLM technologies to learn how to use the technology, what to expect from it, and what kind of custom QA layers are required on top

WHAT MAKES YOU THE PERFECT CANDIDATE

• 2 to 5+ years experience (depending on seniority), or a comparable industry career, with machine learning, natural language processing, data mining, or statistical modeling
• 2 to 5+ years of working experience in Python and the PyData stack or other numerical programming languages
• Experience with analyzing and producing insights from digital product datasets using both qualitative and quantitative techniques
• Strong theoretical background in at least a few among natural language processing (especially modern language models), high-dimensional classifiers, regression models, clustering algorithms, recommender systems, time-series analysis, Bayesian inference, text analytics, knowledge graphs, representation learning (embeddings), computer vision, or social network analysis
• At least some of the data analysis and visualization tools such as pandas, dask, vaex, matplotlib, seaborn, plotly, dash, bokeh, shap, streamlit
• Strong English writing skills to find concise and elegant ways to express difficult concepts in a way that the LLM can understand and act upon
• Educational domain knowledge to be able to rapidly assess the correctness and quality of generated answers
• Ability to develop new evaluation methodologies to assess the quality of prompts in the Brainly domain that have little to do with how traditional ML systems are evaluated
• Motivation to focus full-time on developing hands-on experience and “get in syntony” with each model in order to learn/guess what to expect from each of them in each different context and prompt scenario before even trying

WHAT WILL BLOW OUR MINDS

• Experience with transformers or other deep learning models in production
• Experience with text mining and text analytics
• Experience with data engineering, ETL jobs, or data streaming applications
• Knowledge of at least some of the data engineering technologies such as Spark, DataBricks, Glue, EMR, Docker, Kubernetes, SQL, key-value stores, Redshift, Snowflake
• Familiarity with at least some of the ML technologies such as AWS SageMaker, Tensorflow Extended, PyTorch, Spark ML, scikit-learn, XGBoost, KubeFlow, MLFlow, or related frameworks

WHAT YOU GET BY JOINING BRAINLY

• We want to see you grow along with us – you will have 800$ per year for personal development, extra time for attending conferences and workshops, and unlimited access to an online learning platform (courses from Coursera, Udacity, Udemy, Busuu, Harvard ManageMentor, and many others!)
• Health is important, which is why at Brainly, we fully cover private health & dental care packages for you and your family and provide you with a sport card (Multisport Plus)
• You will also get an access to online individual psychological consultations with professionals in English, Polish & Ukrainian via the Mental Health Helpline
• Flexible working hours – working requires communication, so we work within the European business hours, but we also know that life may be unpredictable, so if you need to jump out of work (doctor’s appointment, emergency, anything) – no problem!
• Your personal concierge AskHenry will support you in your daily duties, eg. planning your dream vacation
• You can join internal communities and contribute to charity, diversity and inclusion initiatives, take part in great internal events or represent Brainly at conferences or meet-ups
• We also provide stock options

WHAT IS BRAINLY

Brainly is a leading learning platform worldwide with the most extensive Knowledge Base for all school subjects and grades. Hundreds of millions of students, parents and educators rely on Brainly as the proven platform to accelerate understanding and learning. Based in Kraków, Poland, with offices in New York City, and Barcelona, Brainly apps and websites are visited by users from over 35 countries. Backed by Prosus, Point Nine Capital, General Catalyst, Runa Capital, Learn Capital and Kulczyk Investments.

Learn more about Brainly at www.brainly.com

By sending us your application you agree that Brainly sp. z o.o. will process your personal data to participate in this recruitment process. If you want to know more about how Brainly processes your personal data please check here [https://careers.brainly.com/data-protection].