Machine Learning Engineer

Role Overview

• Scale Addi’s competitive advantage by building a world-class ML Ops foundation that accelerates the transition from model prototype to production, while ensuring our AI systems from credit scoring to generative agents are resilient, cost-efficient, and seamlessly integrated into our core financial product.
• Ensure ML/AI systems can be served reliably in production, maintaining strong operational excellence for availability, latency, and incident response, in partnership with the Data Scientist role for model/agent logic and iteration.
• Build and maintain the serving and integration layer for ML/AI solutions (APIs, connectors, asynchronous execution patterns), enabling seamless integration with internal systems and Ops tooling.
• Establish clear mechanisms for monitoring and reliability of ML/AI systems in production (dashboards, alerts, core KPIs, regression detection, and data/feature quality checks).
• Enable repeatable delivery for ML/AI services through strong engineering practices (CI/CD, testing, release strategies, rollback, and operational runbooks).
• Make contributions to our Architecture Decision Records repository by evaluating and proposing platform upgrades for ML/AI systems (e.g., feature serving patterns, workflow orchestration, scalable storage) to improve reliability, scalability, and reuse.

Requirements

• Proven experience in architecting and serving production-grade ML systems
• 4–7 years of experience in software engineering, with at least 3 years focused specifically on ML Ops or Data Engineering in a production environment
• Demonstrates the ability to design high-availability serving layers using APIs (FastAPI, gRPC) and asynchronous execution patterns to handle high-concurrency fintech workloads
• Possesses a deep understanding of the "handshake" between data science and engineering, ensuring models are packaged, versioned, and integrated into internal systems without friction
• Expert-level knowledge of AWS (or similar), Kubernetes, Airflow/Prefect, and Databricks/Spark
• Track record of implementing request batching and model quantization to balance high-performance throughput with infrastructure costs
• Possesses strong technical fluency in the Python and Data ecosystem
• Exhibits advanced Python engineering skills, moving beyond simple scripting to build modular, testable, and maintainable codebases
• Expert-level knowledge of core ML libraries (NumPy, Pandas, scikit-learn) and at least one deep learning framework (PyTorch or TensorFlow)
• Solid expertise in data-intensive stacks like Spark or Databricks and the ability to write complex, optimized SQL for feature extraction and data validation
• Experienced in establishing mission-critical observability and reliability
• Has a demonstrated ability to build comprehensive monitoring suites (logs, metrics, traces) that detect not just system downtime, but ML-specific failures like data drift or feature quality regressions
• Track record of leading incident response and post-mortems, with a focus on reducing Mean Time to Recovery (MTTR) for model-related production issues
• Proven ability to implement automated alerting and regression detection that prevents degraded models from impacting the end-customer experience
• Demonstrates a mastery of ML orchestration and engineering best practices
• Proven experience in building repeatable CI/CD pipelines for ML (MLOps), including automated testing, canary releases, and seamless rollback strategies
• Has solid expertise in workflow orchestration tools (e.g., Airflow, Prefect) and storage patterns (Postgres, Vector DBs) required for complex ML lifecycles
• Experienced in contributing to Architecture Decision Records (ADRs) to standardize feature serving patterns and scalable storage across the engineering org
• Track record of building and scaling AI Agentic systems
• Possesses practical experience with the components of modern AI agents, including RAG (Retrieval-Augmented Generation), orchestration frameworks (LangChain/LlamaIndex), and guardrail implementation
• Demonstrates an understanding of the unique operational challenges of LLMs, such as token cost management, prompt versioning, and latency optimization
• Experienced in evaluating and integrating graph-based architectures or graph databases when required for complex data relationship mapping
• Exhibits exceptional cross-functional communication and ownership
• Proven ability to translate highly technical infrastructure bottlenecks into clear business risks or opportunities for non-technical stakeholders
• Demonstrates an "Ownership Mentality" by taking end-to-end responsibility for the reliability of the ML platform, from the initial architectural proposal to 2:00 AM incident resolution
• Varies communication style effectively to mentor Data Scientists on engineering best practices while collaborating with Product Managers on roadmap feasibility.

Tech Stack

• Airflow
• AWS
• GRPC
• Kubernetes
• Numpy
• Pandas
• Postgres
• Python
• PyTorch
• Scikit-Learn
• Spark
• SQL
• Tensorflow

Benefits

• Work on a problem that truly matters
• Be part of something big from the ground up
• Unparalleled growth opportunity
• Join a world-class team
• Competitive compensation & meaningful ownership