Senior MLOps/DevOps Engineer

About Yubi

Yubi, formerly known as CredAvenue, is re-defining global debt markets by freeing the flow of finance between borrowers, lenders, and investors. We are the world's possibility platform for the discovery, investment, fulfillment, and collection of any debt solution. At Yubi, opportunities are plenty and we equip you with tools to seize it.

In March 2022, we became India's fastest fintech and most impactful startup to join the unicorn club with a Series B fundraising round of $137 million.

In 2020, we began our journey with a vision of transforming and deepening the global institutional debt market through technology. Our two-sided debt marketplace helps institutional and HNI investors find the widest network of corporate borrowers and debt products on one side and helps corporates to discover investors and access debt capital efficiently on the other side. Switching between platforms is easy, which means investors can lend, invest and trade bonds - all in one place. All of our platforms shake up the traditional debt ecosystem and offer new ways of digital finance.

Yubi Credit Marketplace - With the largest selection of lenders on one platform, our credit marketplace helps enterprises partner with lenders of their choice for any and all capital requirements.

Aspero - Fixed income securities platform for wealth managers & financial advisors to channel client investments in fixed income

Financial Services Platform - Designed for financial institutions to manage co-lending partnerships & asset based securitization

Spocto - Debt recovery & risk mitigation platform

Accumn - Dedicated SaaS solutions platform powered by Decision-grade data, Analytics, Pattern Identifications, Early Warning Signals and Predictions to Lenders, Investors and Business Enterprises

So far, we have on-boarded over 17000+ enterprises, 6200+ investors & lenders and have facilitated debt volumes of over INR 1,40,000 crore.

Backed by marquee investors like Insight Partners, B Capital Group, Dragoneer, Sequoia Capital, LightSpeed and Lightrock, we are the only-of-its-kind debt platform globally, revolutionizing the segment.

At Yubi, People are at the core of the business and our most valuable assets. Yubi is constantly growing, with 1000+ like-minded individuals today, who are changing the way people perceive debt. We are a fun bunch who are highly motivated and driven to create a purposeful impact. Come, join the club to be a part of our epic growth story.

Responsibilities

• Develop and maintain

  the core ML platform to standardize model development and deployment workflows.

• Create and evangelize

  reusable MLOps components to accelerate the Data Science model lifecycle.

• Design and implement

  seamless integration of trained models with diverse production systems and products.
• Implement robust logging and instrumentation for real-time monitoring of model scoring requests in production.

• Establish automated systems

  for continuous model monitoring and performance-based retraining.

• Design and build advanced deployment strategies

  , including

  Blue-Green Deployment

  and A/B testing frameworks that support canary and shadow models.

• Optimize model inference at scale:

  Implement

  request-based scaling (e.g., KEDA)

  , ensure

  eventual distribution of requests to pods

  , and explore removing node affinity for flexible load distribution.

• Enhance CI/CD and deployment efficiency:

• Reduce container startup time (e.g., Docker layer optimization).

• Integrate unit test case flow

  directly into the deployment pipeline.

• Enable auto deployment from GitHub

  for seamless, GitOps-style operations.

• Scale and manage compute resources:

• Implement

  multi-resource configuration

  for optimal pod provisioning.
• Troubleshoot and fix issues related to

  KEDA-based scaling in QA and Production environments.

• Explore advanced GPU architectures (e.g., NVIDIA MIG)

  and implement

  fractional GPU

  usage for cost and resource optimization.

• Automate Model and Infrastructure Configuration:

• Implement

  automatic model loading from MLflow

  using cloud storage (EFS/S3) for efficient serving.
• Investigate and implement

  Agentic auto-configuration of pods

  to intelligently adapt resources based on traffic/load.
• Integrate and optimize data pipelines necessary for scheduled model retraining and production updates.
• Evaluate and integrate optimal open-source frameworks and proprietary tools into the MLOps pipeline.

Experience & Expertise :

• 3+ years of professional experience

  in

  MLOps, software engineering,

  and successfully deploying production-ready machine learning models.

• Strong expertise in Python

  and advanced scripting for pipeline automation and tooling development.

• Extensive hands-on experience

  with

  Docker and Kubernetes

  for building, managing, and debugging containerized ML environments and scalable production deployments.
• Deep proficiency with

  MLOps frameworks

  (e.g.,

  MLflow, Seldon, Kubeflow

  ) and production-level

  cloud ML services

  (e.g., AWS SageMaker, Azure ML, GCP AI Platform).
• In-depth understanding of

  public cloud infrastructure

  and services (

  AWS, Azure, or GCP

  ).
• Proven ability to

  evaluate, prototype, and implement

  open-source tools to define and guide the MLOps technology stack.
• Excellent

  problem-solving skills

  and a strong analytical mindset.