SDE III - Machine Learning

We are looking for an SDE 3 - Machine Learning to lead the technical architecture for Glance's AI shopping agent. You will own the core engineering for our ML Inference Platform, Evaluation Harnesses, and rapid 01 prototyping frameworks. In this role, you will set technical standards, design reliable distributed ML systems, and work across teams to scale complex agentic workflows.

Responsibilities:

• Inference Platform Architecture: Architect low-latency recommendation inference systems, distributed vector search, feature store integrations, and real-time streaming updates.
• Evaluation Infrastructure: Build comprehensive offline and online evaluation pipelines, golden dataset testing tools, tool-call tracking, and automated regression suites.
• Agentic Core and New Initiatives: Lead the engineering execution for rapid 01 prototyping of AI shopping agent features, building flexible orchestration runtimes, tool registries, and gateway middleware.
• Technical Leadership: Partner with Applied Science, Product, and Data Platform teams to translate complex research concepts into clean, production-grade systems.

Requirements:

• System Architecture: Track record of designing, building, and scaling distributed ML systems and production serving pipelines.
• Deep ML Systems Knowledge: Experience with model optimisation (quantisation, TensorRT, and torch.compile ), vector search indexing, and real-time feature delivery.
• Agent and LLM Infrastructure: Understanding of LLM orchestration, agentic tool-use harnesses, and multimodal model serving.
• Technical Influence: Ability to mentor engineers, guide design reviews, and raise the engineering bar across teams.

You might thrive in this role if you have:

• BTech or MTech in Computer Science, Machine Learning, or a related quantitative field.
• 6+ years of software engineering experience, with 5+ years dedicated to ML platforms, inference engines, or recommendation systems.
• Expertise in building high-throughput, low-latency recommendation inference, vector search, and feature serving infrastructure.
• Proficiency in Python or Java and production ML tools across PyTorch, Triton, and distributed frameworks (Ray/Spark).
• Experience architecting automated evaluation harness suites and production monitoring for ML systems.
• Strong communication skills, able to simplify complex systems and align stakeholders across product and engineering.
• Self-directed, curious, and comfortable driving ambiguous 01 technical initiatives.