Senior ML/AI Engineer (Proctoring)

Company Description

Svitla Systems, Inc. is a global digital solutions company with over 20 years of experience, crafting more than 5,000 transformative solutions for clients worldwide. Our mission is to leverage digital, cloud, data, and intelligent technologies to create sustainable solutions for our clients, enhancing their growth and competitive edge. With a diverse team of over 1,000 technology professionals, Svitla serves a range of clients from innovative startups to Fortune 500 companies across 20+ industries. Svitla operates from 10 delivery centers globally, specializing in areas like cloud migration, data analytics, web and mobile development, and more. We are proud to be a WBENC-certified business and one of the largest, fastest-growing women-owned IT companies in the US.

Role Description

This is a fully remote, full-time, long-term contractual position with one of our clients who is building the next generation of secure, real-time proctoring solutions for high-stakes exams. We’re looking for a Senior ML/AI Engineer to architect, implement, and maintain Azure-based AI models that power speech-to-text, computer vision, identity verification, and intelligent chat features during exam sessions.

Responsibilities

- Implement real-time speech-to-text transcription and audio-quality analysis using Azure AI Speech.  

- Build prohibited-item detection, OCR, and face-analysis pipelines with Azure AI Vision.  

- Integrate Azure Bot Service for rule-based, intelligent chat support.  

- Collaborate with our DevOps Engineer on CI/CD and infrastructure-as-code for AI model deployment.  

- Train, evaluate, and deploy object-detection models (e.g., screen-reflection, background faces, ID checks) using Azure Custom Vision.  

- Develop and maintain skeletal-tracking models (OpenPose/MediaPipe) for gaze-anomaly detection.  

- Fine-tune Azure Face API for ID-to-headshot matching at session start and continuous identity validation.  

- Expose inference results via REST APIs in partnership with backend developers to drive real-time proctor dashboards and post-session reports.  

- Monitor model performance, manage versioning/retraining workflows, and optimize accuracy for edge-case scenarios.

Qualifications

- Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, or related field.  

- 5+ years of professional ML/AI experience, with at least 2 years working on production-grade Azure Cognitive Services.  

- Strong Python skills plus 3+ years with TensorFlow or PyTorch.  

- Hands-on experience (1–2 years) with:  

 - Azure AI Speech (speech-to-text, audio analysis)  

 - Azure AI Vision (object detection, OCR, face analysis)  

 - Azure Custom Vision model training and deployment  

 - Azure Face API fine-tuning and biometric matching  

 - Azure Bot Service integration  

- Solid understanding of CI/CD practices and tools (Azure DevOps, Docker, Kubernetes) with 2+ years of collaboration on AI model deployments.  

- 2+ years building and consuming RESTful or gRPC APIs for AI inference.  

- Proven track record of monitoring and optimizing model performance in production.

Good to haves

- 1+ year with skeletal-tracking frameworks (OpenPose, MediaPipe).  

- Familiarity with Azure ML Studio, ML pipelines, or MLflow for model versioning and retraining.  

- Experience with edge-deployment frameworks (TensorFlow Lite, ONNX Runtime).  

- Background in security and compliance for biometric data (GDPR, PCI-DSS).  

- Azure AI Engineer Associate or Azure Data Scientist Associate certification.

Additional Information

-The role is a Fully Remote, Full-Time, Long-Term, Contractual position

-The hiring process includes an initial screening by the recruitment team, an HR motivation interview, an internal tech screening, a client technical interview and finally the client management interview

-The salary range for this position is 50-70 LPA (INR)

-The position needs to be filled on priority, only candidates who have an official notice period (or remaining time on their notice period) of <=30 days will be screened