Jaydeep Vagh

Temporal-based scheduled microservice for computing multi-window SLO error-budget burn rates

Cost anomaly alerts to PostgreSQL history logs, Slack workspace channels, and PagerDuty endpoints.

The latency-engine is a high-performance Kafka consumer worker service

Latency Baseline Worker

Proactive LLM cost intelligence powered by Google TimesFM.

LLM observability SDK with OTEL tracing, auto-instrumentation, and REST management API.

Go tracer

Code Review Buddy MCP Server

utility tool

Production-ready multi-label toxicity classification worker service

NLI Worker

Quality Baseline Worker

Perplexity scoring microservice

A responsive, high-fidelity browser client for 1-on-1 audio/video calling

A clean, high-performance WebSocket signaling server for 1-on-1 audio/video calling

Production-ready multi-label toxicity classification service

Did the LLM make things up, or stay grounded in context?

Semantic-coherence computes cosine similarity It answers the coherence question

Kafka consumer for LLM span cost aggregation and retroactive token budget reconciliation

Temporal worker package for scheduled EWMA baseline updates and cost anomaly detection.

The budget-provisioner is an internal FastAPI REST microservice

LLM observability SDK with OTEL tracing, auto-instrumentation, and REST management API.

Go AI orchestrator using Cloudflare LLMs with stateful semantic memory.

Contract-driven embedding worker with pluggable providers, schema validation, and REST API.