mkurman/zorai - Page 8

chainladder

Property & casualty insurance loss reserving in Python. Chain ladder, Bornhuetter-Ferguson, Cape Cod, bootstrap simulation, and loss development pattern estimation. Actuarial triangle operations.

clip

OpenAI CLIP — contrastive language-image pre-training. Zero-shot image classification, image-text similarity, concept search, and cross-modal retrieval. Embed images and text into shared space.

crewai

CrewAI — multi-agent AI framework. Role-based agents with defined goals, tools, and memory. Hierarchical and sequential task execution. Human input delegation and process orchestration.

darts

Darts — time series forecasting library by Unit8. Unified API across ARIMA, Prophet, CatBoost, N-BEATS, TFT, TCN, Transformer, and RNN models. Backtesting, probabilistic forecasting, and covariate support.

dgl

Deep Graph Library (DGL) — graph neural network framework. GCN, GAT, GraphSAGE, RGCN, and custom message-passing. Heterogeneous graphs, temporal graphs, and large-scale training with mini-batch sampling.

econml

EconML (Microsoft) — heterogeneous treatment effect estimation. Double ML, Causal Forest, Deep IV, and metalearners (S-Learner, T-Learner, X-Learner). Orthogonal learning for causal effects from observational data.

epidemiology

Disease modeling and epidemiological analysis: SIR/SEIR compartmental models, R0 estimation, outbreak simulation, incidence/prevalence forecasting, and intervention impact modeling.

fredapi

Federal Reserve Economic Data (FRED) API client. 800,000+ US and international economic time series: GDP, inflation, unemployment, interest rates, industrial production. Direct data access for macro research.

freqtrade

Open-source crypto trading bot. Strategy development in Python, backtesting, hyperparameter optimization, dry-run and live trading. Supports major exchanges via CCXT. Telegram integration for monitoring.

google-adk

Google Agent Development Kit (ADK). Code-first Python toolkit for building, evaluating, and deploying AI agents. Multi-agent orchestration, tool integration, built-in evaluation, and deployment to Vertex AI.

guardrails-ai

Guardrails AI — LLM output validation and guardrails. Define guardrails as XML/JSON specs, validate outputs against structural and semantic constraints, correct/retry on failure, and audit model behavior.

huggingface-tgi

HuggingFace Text Generation Inference (TGI). High-performance LLM serving with continuous batching, tensor parallelism, watermarking, and OpenAI-compatible API. Native HF model hub integration.

lancedb

LanceDB — serverless vector database for AI. Columnar storage on Lance format, zero-copy access, multimodal search (text + images + audio), and direct DataFrame integration. No separate server.

langgraph

LangGraph — orchestrate LLM agents as stateful graphs. Multi-agent coordination, persistent state, human-in-the-loop, streaming, checkpointing, and conditional control flow. Build complex agent workflows.

lifelines

Survival analysis in Python: Kaplan-Meier, Cox proportional hazard, Aalen additive, parametric models, and competing risks. Censored data handling for churn, clinical, and actuarial applications.

llama-cpp

LLM inference in C/C++ with Python bindings. GPU acceleration via CUDA/Metal/Vulkan, 2-8 bit quantization (GGUF), KV cache, and grammar-based sampling. Run Llama, Mistral, Gemma, Phi locally.

medical-coding

Medical code mapping and classification tools. ICD-10-CM/PCS, CPT, SNOMED CT, HCPCS, LOINC, RxNorm. Code validation, mapping between terminologies, HCC risk adjustment, and reimbursement modeling.

milvus

Milvus — cloud-native vector database for billion-scale similarity search. GPU-accelerated indexing, hybrid search, multi-vector, streaming, and time travel. Distributed deployment with Kubernetes.

mlflow

MLflow — open-source MLOps platform. Experiment tracking, model registry, packaging, deployment, and evaluation. Multi-cloud ML workflows with reproducible runs and artifact logging.

nixtla

Nixtla ecosystem — statsforecast (statistical), neuralforecast (deep learning), hierarchicalforecast, and MLForecast. Production time series forecasting with AutoARIMA, ETS, Theta, Transformers, and ensemble blending.

nvidia-nemo-guardrails

NVIDIA NeMo Guardrails — programmable guardrails for LLM applications. Colang-based dialog management, topical rails (fact-checking, moderation), safety rails, and security rails for production AI.

ollama

Local LLM runner. One-command setup for Llama, Mistral, Gemma, Qwen, DeepSeek, Phi, and 100+ models. OpenAI-compatible API, model management, GPU acceleration, and custom Modelfile creation.

onnx-runtime

ONNX Runtime — cross-platform ML inference optimizer. Convert PyTorch, TensorFlow, scikit-learn models to ONNX. GPU, CPU, and mobile acceleration. Quantization, graph optimization, and custom ops.

open-clip

OpenCLIP — open-source implementation of CLIP trained on LAION-5B/OpenCLIP datasets. Multi-head attention pooling, SigLIP loss variants, and wide model zoo (ViT, ConvNeXt, EVA). Community-driven.

openai-agents-sdk

OpenAI Agents SDK — build agentic workflows with handoffs, guardrails, and tool integration. Single-agent to multi-agent orchestration. Tracing and observability. Python-first SDK from OpenAI.

peft

Parameter-Efficient Fine-Tuning (PEFT) library. LoRA, QLoRA, AdaLoRA, IA3, Prefix Tuning, P-Tuning, Prompt Tuning. Fine-tune large models with minimal memory overhead. Hugging Face ecosystem integration.

plaid-finance

Plaid financial data API: bank accounts, transactions, balances, income, assets, identity verification, and ACH payments. Connect to 12,000+ financial institutions for personal finance and lending apps.

prophet

Meta Prophet — forecasting at scale. Additive model with yearly/weekly/daily seasonality, holiday effects, changepoints, and trend decomposition. Handles missing data and outliers automatically.

pyportfolioopt

Portfolio optimization library: mean-variance, Black-Litterman, CVaR optimization, risk parity, Hierarchical Risk Parity (HRP), and CLA. Factor models, shrinkage estimators, and ex-ante risk analysis.

qdrant

Qdrant — vector similarity search engine. Payload filtering, quantized indexing, multi-tenant, and horizontal scaling. REST and gRPC API. Docker-native deployment for production RAG and recommendation.

riskfolio-lib

Portfolio risk and optimization: mean-variance, risk parity, CVaR, CDaR, worst-case, and robust optimization. Factor models, Black-Litterman, NCO. Supports plotting and interactive dashboards.

sdv

Synthetic Data Vault (SDV) — generate synthetic tabular data. Single-table, multi-table, and sequential data synthesis. CTGAN, TVAE, CopulaGAN, GaussianCopula. Privacy metrics and evaluation.

sglang

Structured Generation Language for LLM serving. RadixAttention prefix caching, constrained decoding (JSON, grammar), OpenAI-compatible API, and multi-turn optimization. Fast inference with structured output guarantees.

speechbrain

SpeechBrain — PyTorch speech toolkit. ASR, speaker recognition, speech separation, diarization, enhancement, language identification, and TTS. Recipe-based training with pre-trained model zoo.

stripe-payments

Stripe payments integration: charges, subscriptions, invoices, webhooks, Connect platform, and fraud prevention (Radar). Build payment workflows, recurring billing, and marketplace payouts.

trl

Transformer Reinforcement Learning library (TRL). Supervised fine-tuning (SFT), reward modeling, PPO, DPO, KTO, GRPO for RLHF. Process reward models and language model alignment.

unsloth

Fast QLoRA/QLoRA fine-tuning with 2x faster training and 50% less memory. Supports Llama, Mistral, Gemma, Qwen, DeepSeek, Phi, Yi, Falcon. Flash Attention, 4-bit quantization. No quality loss.

vllm

Fast LLM inference engine. PagedAttention, continuous batching, tensor parallelism, speculative decoding, and prefix caching. OpenAI-compatible API server. Supports Llama, Mistral, Qwen, DeepSeek, and hundreds of models.

wandb

Weights & Biases — ML experiment tracking and visualization. Log metrics, hyperparameters, model checkpoints, and artifacts. Collaborative dashboards, sweep hyperparameter search, and model registry.

weaviate

Weaviate — open-source vector database with built-in ML. Hybrid search (vector + keyword), generative search, graph connections, multi-modal (text + image), and automatic schema inference.