Middle / Senior ML Engineer

Ищем программиста со знанием OpenCV, YOLO и нейросетевых моделей компьютерного зрения для работы с видео-аналитикой и обучения ИИ-моделей анализа видео.

Нужно разработать модуль анализа видео, который принимает видеопоток с камеры, обнаруживает нужные объекты, отслеживает их движение и формирует результат в виде структурированных данных. Также требуется обучить или дообучить модель под конкретные объекты и сценарии.

Ожидаемый результат: рабочий прототип анализа видео; обученная или дообученная модель YOLO; инструкция по запуску и обучению; скрипты для инференса, тестирования и оценки качества; возможность интеграции результата в основное приложение.

Основные задачи:

• Разработка системы анализа видеопотока с камер.
• Детекция, классификация и трекинг объектов на видео.
• Работа с моделями YOLO: обучение, дообучение, тестирование и оптимизация.
• Подготовка датасетов: сбор, разметка, очистка, аугментация данных.
• Настройка пайплайна обучения модели: train / validation / test.
• Оптимизация модели под работу в реальном времени.
• Интеграция видеоаналитики в существующее приложение или сервер.
• Работа с RTSP/IP-камерами, видеопотоками, файлами и кадрами.
• Вывод результатов анализа: координаты объектов, классы, confidence, события, логи.

Требуемые навыки:

• Python, OpenCV, YOLO: YOLOv5 / YOLOv8 / YOLOv10 / YOLOv11 или аналогичные модели; PyTorch или другой ML-фреймворк; Linux, Git.
• Опыт обучения и дообучения моделей компьютерного зрения.
• Понимание принципов object detection, classification, tracking.
• Работа с видео: RTSP, IP-камеры, FFmpeg / GStreamer будет плюсом.
• Умение работать с датасетами и инструментами разметки.
• Умение давать реалистичную оценку сроков разработки в часах с учетом технических рисков, неопределенностей и этапов реализации.
• Можете в короткие сроки разработать рабочий MVP и последовательно улучшить модель до нужных метрик качества, надежности и производительности.

Желательные навыки:

• Оптимизация моделей: ONNX, TensorRT, CUDA.
• Работа с edge-устройствами: Raspberry Pi, Orange Pi, Jetson, промышленные mini-PC.
• Опыт ускорения инференса на CPU/GPU/NPU.
• ByteTrack, DeepSORT, SORT или другие алгоритмы трекинга.
• Docker.
• FastAPI / Flask / Node.js для интеграции модели в сервер.
• Опыт построения систем видеоаналитики в реальном времени.

Для отклика желательно прислать:

• Примеры проектов с OpenCV / YOLO / Computer Vision.
• GitHub или портфолио.
• Описание опыта обучения моделей.
• Какие модели и фреймворки использовали.
• Какой FPS удавалось получить на видео в реальном времени.

Что мы предлагаем:

• Проектная или постоянная работа.
• На время ИС 5/2 работа на месте производства в п. Мосрентген, после ИС можно также работать из офиса на м. Калужская.
• Оплата обсуждается по результатам собеседования и оценки объема задач.

Why Socure?

Socure is building the identity trust infrastructure for the digital economy — verifying 100% of good identities in real time and stopping fraud before it starts. The mission is big, the problems are complex, and the impact is felt by businesses, governments, and millions of people every day.

We hire people who want that level of responsibility. People who move fast, think critically, act like owners, and care deeply about solving customer problems with precision. If you want predictability or narrow scope, this won’t be your place. If you want to help build the future of identity with a team that holds a high bar for itself — keep reading.

Overview

We are seeking a Staff Data Scientist with deep expertise in computer vision and machine learning to provide technical leadership for our document verification solutions. This role is responsible for setting technical direction, guiding data scientists through complex challenges, and ensuring the quality of ML solutions across the full model lifecycle—from data strategy and model development to deployment, monitoring, and production usage.

You will serve as a domain expert and trusted technical advisor, raising the technical bar across the team while partnering with engineering, product, and customers to deliver effective document verification solutions.

Responsibilities

• Serve as a domain expert in computer vision and machine learning, providing technical direction and guidance across document verification projects.

• Oversee the end-to-end ML lifecycle, including data selection, model development, evaluation, deployment, monitoring, and continuous improvement.

• Guide and mentor data scientists through complex modeling and production challenges, helping teams make sound technical decisions and tradeoffs.

• Establish best practices and technical standards for model development, evaluation, and production readiness.

• Identify technical risks and opportunities across projects and drive solutions that improve model performance, scalability, and reliability.

• Partner with engineering, product, and customers to translate business and customer needs into effective modeling strategies.

• Communicate model behavior, performance, limitations, and technical decisions clearly to both technical and non-technical stakeholders.

Requirements

• Bachelor’s degree in Computer Science or a related field, or equivalent experience, with 10+ years or equivalent professional experience and deep expertise in computer vision and machine learning.

• Proven experience designing, deploying, and improving production ML systems at scale.

• Demonstrated technical leadership across multiple ML projects, including mentoring and guiding other experienced data scientists.

• Strong understanding of the full ML lifecycle, from data strategy and model development through deployment and production monitoring.

• Strong technical judgment and ability to navigate complex and ambiguous modeling problems.

• Excellent written and verbal communication skills, including the ability to influence technical, business, and customer stakeholders.

Nice to Have

• Master’s or Ph.D. in Computer Science, Machine Learning, Computer Vision, or a related field.

• Experience with document verification, OCR, identity, fraud detection, or image forensics.

• Experience with transformer-based vision models, multimodal systems, foundation models, or LLMs.

• Experience working with enterprise customers or in regulated environments.

Socure is an equal opportunity employer that values diversity in all its forms within our company. We do not discriminate based on race, religion, color, national origin, gender, sexual orientation, age, marital status, veteran status, or disability status.
If you need an accommodation during any stage of the application or hiring process—including interview or onboarding support—please reach out to your Socure recruiting partner directly.

Follow Us!

YouTube | LinkedIn | X (Twitter) | Facebook

We are looking for a Junior Research Scientist to work on foundation models for financial data. Several of them already run in production on bureau, transactional and graph data; your work is to extend that line to new data domains and new downstream tasks — credit risk, acquisition, anti-fraud, LTV forecasting.

We work with Python, PyTorch, MLFlow, AWS, Snowflake and modern AI coding tools such as Claude Code, OpenAI Codex and Cursor.

This is a great opportunity for someone at the beginning of their career and who wants to work on large-scale, technically challenging projects. 

Challenges that await you:

• Train self-supervised models on discrete sequences to beat the SOTA and achieve business impact in downstream tasks across Plata, such as: credit risk, transaction anti-fraud, acquisition and LTV forecasting
• Stay on top of SOTA research, applying the latest NLP and DL techniques to fintech models
• Work with large multimodal datasets: tabular, behavioral, transactional, device and network, text, time series, graphs
• Optimize the utilization of compute resources for both training and inference
• Own and develop solutions end-to-end, from idea to data collection to experiments to training runs to inference optimization to evaluating impact
• Perform rigorous evaluations
• Write articles and speak at industry conferences

What makes you a great fit:

• Education: mathematics, engineering, computer science, artificial intelligence or another strong quantitative field
• Strong foundation in machine learning and deep learning, with hands-on experience building and experimenting with models using PyTorch and LLMs
• Understanding of dataset design and data mixing for deep learning training
• Demonstrated understanding of how to maximize the utilization of compute resources and efficiently perform experiments to validate research ideas
• Solid communication skills
• High autonomy and ownership

Your bonus skills:

• Published research in machine learning or deep learning
• Personal projects or startup experience in AI, machine learning or deep learning
• Internship experience in research, ML or DL teams
• B1 or higher English level for effective communication with an international team

Our ways of working:

• Innovative Spirit: A commitment to creativity and groundbreaking solutions
• Honest Feedback: valuing open, transparent communication
• Supportive Team: a strong, collaborative community
• Celebrating Achievements: recognizing our wins together
• High-Tech Environment: a team full of smart and revolutionary people who dare to challenge the status quo of incumbent finances 

Our benefits:

• Relocation support to one of our hubs — Serbia, Georgia or Mexico — with assistance for the employee and their family
• Flexible work from one of our offices or remotely within time zones from GMT-2 to GMT+5
• Healthcare Coverage
• Education Budget: Language lessons, professional training and certifications
• Wellness Budget: Mental health and fitness activity reimbursements
• Vacation policy: 20 days of annual leave and paid sick leave