akiva ai

All

10 repositories

toolkit-eval-harness
Public
AI/ML evaluation harness -- versioned golden test suites, deterministic scoring, regression detection, and CI/CD-integrated reporting.
python benchmarking evaluation
python benchmarking evaluation ci-cd regression-testing ml-testing
Python
•
MIT License
•0•1•0•4•Updated Mar 9, 2026Mar 9, 2026
toolkit-policy-test-bench
Public
Red-team and compliance test harness for LLM outputs -- detects PII leakage, secret exposure, and policy violations with CI-friendly reporting.
python compliance red-teaming
python compliance red-teaming pii-detection policy-testing llm-security
Python
•
MIT License
•0•1•0•4•Updated Mar 9, 2026Mar 9, 2026
toolkit-inference-mesh
Public
Distributed LLM inference mesh for heterogeneous clusters -- pipeline-parallel sharding with SGLang, MLX-LM, and P2P networking. Fork of Parallax.
distributed-computing inference mlx
distributed-computing inference mlx llm-serving model-sharding sglang
Python
•
Apache License 2.0
•0•1•0•14•Updated Mar 9, 2026Mar 9, 2026
toolkit-ml-provenance
Public
ML provenance and SBOM generator -- deterministic manifests with integrity verification and optional cryptographic signing for datasets, configs, code, and mode…
python ml provenance
python ml provenance integrity reproducibility sbom
Python
•
MIT License
•0•1•0•3•Updated Mar 9, 2026Mar 9, 2026
enterprise-crypto
Public
Open-source multi-agent crypto trading system with institutional-grade risk controls, multi-exchange support, and autonomous strategy execution.
open-source typescript crypto
open-source typescript crypto trading multi-agent risk-management
TypeScript
•
Other
•0•1•0•24•Updated Mar 9, 2026Mar 9, 2026
toolkit-cost-optimizer
Public
LLM inference cost and latency optimizer -- log analysis, SLO-based model recommendations, and routing policy simulation.
python latency routing
python latency routing slo cost-optimization llm-ops
Python
•
MIT License
•0•1•0•3•Updated Mar 9, 2026Mar 9, 2026
toolkit-llm-gateway
Public
Enterprise LLM proxy with unified multi-provider API, cost attribution, rate limiting, caching, and analytics -- based on LiteLLM.
proxy gateway openai
proxy gateway openai llm cost-tracking litellm
Python
•
Other
•0•1•0•6•Updated Mar 9, 2026Mar 9, 2026
toolkit-data-contracts
Public
Data contract management and drift detection for ML/LLM pipelines -- automatic schema inference, validation, and statistical profiling with CI/CD integration.
python schema-validation data-quality
python schema-validation data-quality drift-detection ml-pipelines data-contracts
Python
•
MIT License
•0•1•0•4•Updated Mar 9, 2026Mar 9, 2026
toolkit-mmqa
Public
Multimodal dataset QA -- directory scanning, file hashing, and exact duplicate detection for text, image, and audio datasets.
python hashing deduplication
python hashing deduplication data-quality multimodal dataset-qa
Python
•
MIT License
•0•1•0•3•Updated Mar 9, 2026Mar 9, 2026
toolkit-rag-quality
Public
Deterministic RAG evaluation toolkit -- retrieval metrics (recall, precision, MRR), corpus overlap detection, and CI regression gating without model calls.
python information-retrieval ci-cd
python information-retrieval ci-cd data-quality rag retrieval-evaluation
Python
•
MIT License
•0•1•0•3•Updated Mar 9, 2026Mar 9, 2026