SENIOR ML/AI ENGINEER

Location

Remote restrictions apply

See all remote locations

Salary Estimate

N/A

Seniority

Senior

Tech stacks

Azure

Python

+19

Permanent role

9 days ago

Apply now

_Svitla Systems Inc. is looking for a Senior ML/AI Engineer for a full-time position (40 hours per week) in Poland.

_Our client is a leading online training provider. The company offers 7,000+ regulated and non-regulated courses for career development and has 11+ million learners in 67 countries.

The robust e-learning platform supports multimedia content, quizzes, and assessments. Courses are accessible on various devices, including smartphones and tablets. There are regular updates to course content to reflect the latest industry standards and regulations. It is a reliable and reputable company that provides various online training and certification programs across multiple industries, such as real estate, food and beverage, environmental health and safety, and more.

Established in 1997 and headquartered in Austin, Texas, USA, they have earned accreditation from esteemed organizations and regulatory bodies like the International Accreditors for Continuing Education and Training (IACET), ANSI National Accreditation Board (ANAB), and the Texas Real Estate Commission (TREC), among others.

ⓘ View potential timezone restrictions

_Note* – In this project, you will work with a team located entirely in the EU time zone. Occasionally, there may be late meetings with the product owner and client stakeholders due to time zone differences. The client’s managers are aware of this, so calls after 7 PM CET will be rare. However, in critical cases, you should be available to join such meetings, which are not part of the regular daily routine.

_Project Overview

Domain: AI-powered online proctoring for remote examinations
Goal: Develop a high-availability, scalable platform to replace an outdated legacy system
Team: Cross-functional Agile team including AI/ML Engineer, Solution Architect, Backend Developer, and DevOps Engineer
Collaboration: Work closely with technical stakeholders to integrate AI capabilities into a secure and compliant platform
Key Responsibilities: Identity verification, Gaze tracking, Environment validation, Speech analysis, Behavioral anomaly detection; Design, implement, and optimize AI services and models

Technology Stack

Programming Language: Python
ML Frameworks: PyTorch, TensorFlow, Scikit-learn
Cloud Platform: Microsoft Azure
Azure Services: Azure AI Speech, Azure Vision, Azure Custom Vision, Azure Machine Learning Studio, Azure Face API, Azure Bot Service, Azure Cognitive Services

Requirements

4+ years of experience in ML/AI engineering, preferably in Computer Vision, Speech Processing, or Biometric Systems.
Practical experience with Azure AI Speech, Azure Vision & Face API, Azure Custom Vision, Azure Machine Learning Studio
Proficiency in Python and popular ML frameworks (e.g., PyTorch, TensorFlow, Scikit-learn)
Experience deploying or working with OpenPose or other keypoint detection libraries
Familiarity with image/video/audio preprocessing, real-time inference pipelines, and lightweight model optimization
Understanding of the risks and ethics involved in facial recognition and behavioral tracking
Strong problem-solving and debugging skills, with a proactive mindset
Advanced English (written and verbal) for daily communication
Working Hours: 8 hours per day, flexible schedule
*Time Overlap: Required overlap with the team until 19:00 CET.

Responsibilities

Use Azure AI Speech for real-time speech-to-text transcription and audio quality analysis during exam sessions.
Apply Azure AI Vision to enable prohibited item detection, OCR, and face analysis.
Support Azure Bot Service integration for intelligent, rule-based chat communication.
Work with the DevOps Engineer to streamline AI model deployment and infrastructure setup.
Train, evaluate, and deploy object detection models using Azure Custom Vision (e.g., screen reflection, background faces, ID inspection).
Maintain skeletal tracking models for analyzing head/eye movement, enabling gaze anomaly detection and timeline flagging.
Fine-tune Face API to perform ID-to-headshot matching at session start and enable continuous identity validation throughout the exam.
Collaborate with backend developers to deliver AI inference results via APIs, powering real-time proctor dashboards and session reports.
Monitor model performance, manage versioning and retraining workflows, and optimize for accuracy and reliability in edge cases.

We offer