MLOps Engineer (Triton + GPU + Production AI)

Job Description – MLOps Engineer (Triton + GPU + Production AI)

Immediate joining.
Employment Type: Full-time
Project: OTRAS – Next-Gen AI-based Government Exam & Recruitment Platform

MLOps Engineer (Triton + GPU + Production AI)

Role: MLOps Engineer

Experience: 5–10 Years

Location: Andhra Pradesh

Salary: ₹1,00,000 – ₹1,50,000 per month

About the Role

We are building OTRAS, India’s largest next-gen AI-based examination platform serving 250M+ candidates per year.

We need an experienced MLOps Engineer who can productionize large AI/ML models (OMR, OCR, face recognition, fraud detection) using NVIDIA Triton, ONNX, TensorRT, and GPU pipelines.

You will be responsible for deploying, scaling, monitoring, and optimizing AI workloads in a distributed Kubernetes environment.

Key Responsibilities

Model Deployment & Serving

• Deploy PyTorch/TensorFlow models on NVIDIA Triton Inference Server
• Convert models to ONNX and optimize using TensorRT
• Implement batching, dynamic batching, and GPU scheduling
• Build scalable inference APIs (HTTP/gRPC)

Infrastructure & Automation

• Deploy and manage AI workloads on Kubernetes (GPU node pools)
• Automate model CI/CD using GitHub Actions + ArgoCD
• Setup model versioning, canary deployments, and rollback workflows
• Manage the Triton model repository & configs

Monitoring & Optimization

• Implement inference metrics (latency, TPS, GPU utilization)
• Setup monitoring using Prometheus + Grafana
• Optimize inference speed and memory with TensorRT
• Run load tests for 10M+ inference events

Data & Pipelines

• Build ETL workflows for AI datasets
• Automate dataset cleaning, preprocessing
• Integrate with ClickHouse / S3 storage
• Create pipelines for:
• ✔ OMR data ingestion✔ ID card OCR✔ Face detection & liveness scoringSecurity & Reliability
• Ensure secure model access (token-based + mTLS)
• Handle production failures, logs, distributed tracing
• Implement AI/ML model audit trails

Required Skills

• 4+ years experience in MLOps or ML Engineering
• Strong hands-on with:
• ✔ NVIDIA Triton Inference Server✔ ONNX / ONNX Runtime✔ TensorRT✔ PyTorch or TensorFlow✔ CUDA (basic understanding)
• Strong in Docker & Kubernetes

• Experience with CI/CD
• Knowledge of GPU scaling, batching, and memory optimization
• Experience working with large-scale ML systemsBonus Skills

• Experience with Airflow or Kubeflow
• Experience with model quantization
• Familiarity with computer vision
• Knowledge of message queues (Kafka)
• Worked on AI for ID verification / OMR / OCR

Why Join OTRAS?

• Build India’s first AI-powered exam infrastructure
• Work with Go microservices + Kubernetes + Triton
• Massive impact (250M candidates)
• Fast-moving, high-performance engineering culture
• High visibility role with strong growth

Job Types: Full-time, Permanent, Volunteer

Pay: ₹180,000.00 - ₹1,080,070.03 per year

Benefits:

• Health insurance
• Life insurance
• Provident Fund

Ability to commute/relocate:

• Guntur, Andhra Pradesh: Reliably commute or planning to relocate before starting work (Required)

Work Location: In person