Senior Machine Learning Engineer

About the Role

We are seeking an experienced Senior Machine Learning Engineer to lead the development of AI-powered services for our global online learning platform. This position focuses on creating and maintaining the machine learning models that enable real-time speech practice, conversational interaction, and personalized feedback for learners. The role entails end-to-end ownership of various models—including classical, fine-tuned small transformer, edge, and cloud-based models—as well as integrating with external AI providers. This is an active coding position, where you’ll write, maintain, and improve ML services deployed in production, using AI-driven coding tools as part of your daily work.

Responsibilities

• Design, build, and evaluate machine learning models throughout the entire ML lifecycle, balancing traditional statistical methods with fine-tuning of compact transformer-based models optimized for edge devices.
• Develop and manage integrations with hosted large language model (LLM) providers, including prompt design, evaluation processes, multi-provider routing, failover strategies, and cost-latency optimization.
• Implement and operate ML models both on-device and in cloud environments, making informed decisions on deployment based on latency requirements, privacy, and cost considerations.
• Engage with multi-modal perception pipelines spanning speech recognition (ASR), text processing (NLP), and computer vision (CV) according to product needs, applying sound judgment to evaluate trade-offs across modalities.
• Conduct practical production engineering tasks involving setting up, maintaining, and upgrading ML services with focus on service constraints, monitoring, drift detection, and correlating offline model metrics with real user outcomes.
• Build and maintain robust data extraction, transformation, and loading (ETL) pipelines to process learner data from storage and databases, ensuring cleanliness and reduced noise in the input datasets.

Required Skills & Qualifications

• Minimum of 5 years of experience designing, developing, and deploying machine learning systems in production environments, encompassing both classical ML and applied deep learning techniques.
• Proficiency in systems programming languages such as Rust or Go is preferred; strong C++ skills are also beneficial. Python or Julia used for data science, modeling, and exploratory data analysis.
• Hands-on experience with at least one major ML framework (e.g., PyTorch, TensorFlow, JAX), emphasizing real-world production usage over any particular framework allegiance.
• Familiar with Git version control and collaborative development practices including code review processes.
• Daily usage of AI-based coding tools and agents integrated within your development workflow, not as occasional helpers.
• Demonstrated experience integrating hosted LLM services into production platforms, covering prompt engineering, evaluation strategy, cost and latency tradeoffs, and multi-provider orchestration.
• Capability to independently develop models from first principles and fine-tune pretrained models, making rigorous decisions about which approach best fits project constraints.
• Strong communication skills, with ability to articulate complex modeling choices to technical peers and explain product impacts clearly to non-technical stakeholders.

Preferred Qualifications

• Experience with model optimization techniques such as compression, quantization, or distillation for deploying on edge devices.
• Some hands-on exposure in at least two domains among natural language processing (NLP), speech processing (ASR/TTS), or computer vision (CV), with no expectation to possess full expertise in all three.
• Proficiency in managing data ETL workflows involving object storage systems (e.g., S3), relational databases (such as Postgres), and data cleaning or noise reduction methods.

Why Join Us?

• Enjoy broad, full-stack exposure across speech, text, and computer vision within a compact, multidisciplinary team rather than a narrow specialization.
• Contribute creatively to building model-serving architecture from the ground up without legacy systems or imposed technical debts.
• Engage in meaningful engineering decisions emphasizing real-world trade-offs around edge computing and multi-provider LLM integration, not just model accuracy metrics.