Senior Computer Vision Engineer / Старший инженер по компьютерному зрению

Social Discovery Group (SDG) is the 3rd largest social discovery company in the world, uniting 60+ brands with 500 million users. We solve the problems of loneliness, isolation, and disconnection by transforming virtual intimacy into the new normal. Our portfolio includes online communication platforms focusing on AI, game mechanics, and video streaming - Dating.com, DateMyAge, Cupid Media, Dil Mil, Kiseki, and others.

We bring together a team of like-minded people and IT professionals specializing in the creation and development of globally impactful social discovery products. Our international team of 1200 professionals and digital nomads works all over the world.

We're looking for a Senior/Lead Computer Vision Engineer

Your main tasks will be:
- Train and fine-tune generative models such as text-to-image, video-to-image, image-to-image, and IP-adapters;
- Train image classifiers, rankers, and image/video captioning models;
- Develop a CV roadmap for the ML team, keeping up with the latest technologies and advancements in open-source solutions;
- Establish a process for iterative improvement and validation of CV model quality;
- Optionally, train multimodal bots;
- Lead a CV team of approximately 2-3 members.

We expect from you:
- Proven experience in training diffusion models and modifying their architectures;
- Familiarity with Diffusers and Transformers libraries;
- Hands-on experience with IP-adapters;
- Proficiency in training flow matching;
- Knowledge of diffusion model distillation;
- Experience with DPO and other methods for fine-tuning text-to-image models;
- Proficiency in fine-tuning text-to-video models;
- Experience with fine-tuning CLIP models;
- Work with multimodal large language models (LLMs) is a significant plus;
- Expertise in face swap technologies;
- Bonus: Experience in NLP, particularly in training language models and preparing chatbots.

What do we offer:
- REMOTE OPPORTUNITY to work full time;
- Vacation 28 calendar days per year;
- 7 wellness days per year (time off) that can be used to deal with household issues, to lie down and recover without taking sick leave;
- Bonuses up to $5000 for recommending successful applicants for positions in the company;
- Full payment for professional training, international conferences and meetings;
- Corporate discount for English lessons;
- Health benefits. According to the paychecks, if you are not eligible for corporate medical insurance, the company will compensate you with up to $1,000 gross per year per employee. This money can be spent on self-purchase of health insurance or on doctor’s fees for yourself and close relatives (spouse, children);
- Workplace organization. The company provides all employees with an equipped workplace and all the necessary equipment (table, armchair, wifi, etc.) in our offices or co-working locations. In other locations, the company provides reimbursement of workplace costs up to $1,000 gross once every 3 years, according to the paychecks. This money can be spent on the rent of the co-working room, on equipping the working place at home (desk, chair, Internet, etc.) during those 3 years;
- Internal gamified gratitude system: receive bonuses from colleagues and exchange them for time off, merch, team building activities, massage certificates, etc.

Sounds good? Join us now!

RU

Мы ищем старшего/ведущего инженера по компьютерному зрению.

Основные задачи:
- Обучать и дорабатывать генеративные модели, такие как текст-в-изображение, видео-в-изображение, изображение-в-изображение и IP-адаптеры;
- Обучать классификаторы изображений, ранжировщики и модели для создания подписей к изображениям и видео;
- Разрабатывать дорожную карту по компьютерному зрению для команды машинного обучения, следить за новыми технологиями и развитием open-source решений;
- Внедрять процессы для поэтапного улучшения и проверки качества моделей компьютерного зрения;
- По желанию — обучать мультимодальные боты;
- Руководить командой из примерно 2-3 специалистов по компьютерному зрению.

Ожидаем от вас:
- Подтверждённый опыт обучения диффузионных моделей и изменения их архитектур;
- Знание библиотек Diffusers и Transformers;
- Практический опыт работы с IP-адаптерами;
- Умение обучать модели с использованием flow matching;
- Знание методов дистилляции диффузионных моделей;
- Опыт работы с DPO и другими методами тонкой настройки текст-в-изображение моделей;
- Навыки тонкой настройки текст-в-видео моделей;
- Опыт тонкой настройки моделей CLIP;
- Работа с мультимодальными крупными языковыми моделями (LLMs) будет большим плюсом;
- Опыт работы с технологиями замены лиц (face swap);
- Дополнительно приветствуется опыт в NLP, особенно в обучении языковых моделей и подготовке чатботов.