Machine Learning Engineer

A remote India machine learning role at Coinbase, and at 2 years it is the second most accessible job in today's edition. The team is CX Intelligence, which owns the multi agent platform behind Coinbase chat, the Help Center and agent tooling. The work is building the orchestration layer that manages state transitions, context sharing and intent routing between vendor AI, internal multi agent systems and human agents, and writing production grade Python services around it. Familiarity with LangGraph, LangSmith, Google ADK, Vertex AI or AWS Bedrock is named. Two things to know before applying, both stated in the posting: Coinbase is remote first but not remote only, and expects quarterly in person working sessions it calls surges; and candidates may submit a maximum of 3 applications in any 6 month period, so choose carefully.

Required by the posting: 2+ years of professional experience in machine learning and software engineering, with experience shipping production grade ML services. Hands on experience building with modern AI architectures such as LLMs and deep learning, and familiarity with tools including LangGraph, LangSmith, Google ADK, Vertex AI or AWS Bedrock. Strong proficiency in Python with the ability to write clean, maintainable, well tested production code. Working knowledge in at least one domain such as NLP, information retrieval, computer vision or statistical modelling. Ability to write technical design documents and communicate ML system designs clearly to cross functional stakeholders. Responsible use of generative AI with human oversight to deliver business ready outputs.

The real day to day: building and improving the orchestration layer that manages state transitions, context sharing and intent routing across vendor and internal LLM frameworks in a distributed conversational environment; developing production grade Python services that bridge AI and ML capabilities with customer facing products; driving well scoped ML projects from design through delivery; contributing to system design, coding standards and AI/ML best practices; and participating in design reviews to ensure features meet Coinbase standards for security, compliance and performance.

Team context: the CX Intelligence team sits in Coinbase's Enterprise Applications and Architecture organisation and owns the multi agent platform powering Coinbase chat, the Help Center and agent tooling, working with Conversation Design and CX teams. The role is levelled IC4.

Location and working pattern: the board lists this as Remote in India. The posting states that Coinbase is remote first but not remote only, and that you should expect to get together quarterly for in person working sessions the company calls "surges". Factor that travel in before treating this as a fully remote job.

Compensation: unusually, the posting publishes a specific annual base salary figure for this requisition of ₹4,408,400, excluding equity and bonus. Very few postings on this board state Indian salary at all.

Application limit, stated in the posting: candidates may submit a maximum of 3 applications within a 6 month period at Coinbase. Do not spray applications across their board.

Honest fit guidance: the posting opens by saying the bar is high and the environment intense, and that it is not a place for complacency. Take that as written. But the 2 year requirement is genuine and unusually low for a remote role of this quality, so if you have shipped a production ML service and know the modern agent tooling, this is worth your best effort rather than a quick application.