spendguard-ai: Precision Budget Guard & Circuit Breaker
Zero-dependency, sub-microsecond token budget guard, floating-point spend tracking & latency circuit breakers for OpenAI, LiteLLM, Anthropic & DeepSeek. Published globally on PyPI.
Zero-dependency, sub-microsecond token budget guard, floating-point spend tracking & latency circuit breakers for OpenAI, LiteLLM, Anthropic & DeepSeek. Published globally on PyPI.
Resolved IEEE-754 precision underflow (-2.7e-17) occurring on periodic spend resets across Verification Tokens, Users, and Teams. Implemented unit-of-work queue clamping and Pydantic schema validation.
Fixed Popup lifecycle timing bug approved by Sofia Core Lead @TeodorTaushanov (PR #14258). Resolved TreeItemCustom ARIA setsize bug (#14239) diagnosed with Senior Dev @GDamyanov, ensuring accurate screen reader tree semantics (PR #14259).
Engineered first-class support for max_completion_tokens in Azure OpenAI client and LangChain integration, unblocking OpenAI reasoning models (o1, o3-mini).
Architected 5 standalone knowledge graph connectors for Ansible Infrastructure, Excalidraw Visual Diagrams, Hex.pm BEAM, Crates.io Rust, and DigitalOcean Cloud with automated test suites.
Fetching live projects from GitHub...
Select your preferred language below to view and print the A4 document.
Motivated Computer Science graduate with a strong foundation in software engineering, backend architectures, and AI applications. Passionate about designing robust APIs, integrating intelligent NLP models, and maintaining high software quality. Driven to bridge the gap between AI development and production systems through testing, debugging, and reliable software delivery.