AI Products Developer at Plus Fantech

We are engaged in content analysis and labeling, work with ML models and Yandex Tasks, integrate Alice into new products, and fight pirates. We are looking for a person who will help us integrate advanced AI models into content processing and personalization, taking the user experience to a new level.

Next year, we want Fantech's services to gain a full-fledged AI layer and start providing users with a new, "magical" experience: personalized advice and explanations of why this particular product, service, or content might be interesting to a specific person, offering smart assistants within interfaces, and automatic recaps of complex video content. To achieve this, we will launch new AI projects at the intersection of CV, VLM, thematic AI agents, and generative pipelines based on LLMs enriched with personal and service data, as well as develop a deeper understanding of video. Your task will be to help come up with such features and bring them from rapid experiments to reliable production solutions.

On the path from any beautiful technology to a real product lies a multitude of engineering tasks: data preparation and delivery, computational optimization, automation of steps done manually during the prototyping phase, and using human-in-the-loop. We are looking for a developer who will help bring technologies to production, ensure their reliable operation and scaling, and develop the technological and infrastructural components.

What tasks await you

Working with multimodal data You will process text, images, and audio, using both advanced paid and free models, as well as Yandex's internal developments.

Experiments and prototyping You will design quick experiments for MVP: selecting suitable open-source solutions, testing them, analyzing results, and based on the conclusions, working out scaling paths. It is important to be aware of current approaches and adapt them to our tasks.

Fine-tuning and adapting models Sometimes you will need to fine-tune existing models on our data for specific business tasks. The focus is on building offline pipelines for processing large volumes of data; periodically real-time projects will arise where the key requirements will be the efficiency and robustness of solutions.

More about backend at Yandex — in the channel Yandex for Backend

We expect that you

• Know algorithms and data structures
• Can write good quality code in Python
• Understand MapReduce and SQL
• Can independently communicate with manager-customers, formulate technical requirements, and design data delivery architecture together with developers of related components

Will be a plus

• Actively interested in the development and application of neural networks
• Have used various LLMs and VLMs at work or have done pet projects with their application
• Love movies, music, and books as much as we do