Shrey Satapara enhanced cost-tracking reliability for LiteLLM within the All-Hands-AI/agent-sdk repository by addressing issues with provider-prefixed model names. Using Python for backend development, Shrey implemented an API integration that ensures cost calculations and telemetry reporting accurately reflect model usage, which is essential for analytics and governance. The work involved developing a feature that correctly parses and tracks costs for models with provider prefixes, resolving a bug that previously led to inaccurate reporting. Through careful unit testing and attention to backend logic, Shrey’s contribution improved the transparency and reliability of cost data, supporting better decision-making for platform stakeholders.