stanfordnlp/dspy Python
Stanford's language model programming framework
Open Source
Bug fixes, features, and documentation contributed to ML frameworks and scientific Python libraries. github.com/kvr06-ai
AI & ML Frameworks
langchain-ai/langgraph Python
LangChain's agent orchestration framework
- #7020fix(langgraph): add missing | None to aupdate_state values type
- #7019fix(langgraph): skip caching ERROR/INTERRUPT writes in SyncPregelLoop
- Filed 4 detailed bug reports (#7015-#7018) documenting sync/async parity issues with reproduction steps.
567-labs/instructor Python
Structured outputs for LLMs
Also contributed to
- huggingface/diffusers Fix double-shifted sigma_min/sigma_max in FlowMatchEulerDiscreteScheduler
- emcie-co/parlant Embedding cache eviction regression test + OpenRouter retry policy fix
- agno-agi/agno Fix knowledge_table read from Knowledge's contents_db
- anthropics/claude-agent-sdk-python Document TypedDict runtime dict behavior in types.py
Scientific & Statistical Python
py-why/dowhy Python
Microsoft's causal inference and counterfactual estimation library
shap/shap Python
Game-theoretic feature attribution (Shapley values)
Also contributed to
- statsmodels/statsmodels Fix docstring formula display in SVAR class
- optuna/optuna Correct return type annotations for Axes in visualization functions
- explosion/spaCy Fix memory_zone cleanup on exception
Game Theory & Research Tools
google-deepmind/open_spiel C++/Python
DeepMind's framework for research in games and game theory
Original tools
spielviz PyPI
Python visualization library for OpenSpiel game trees, information sets, and strategy profiles
trust-based-public-goods-game research
Multi-agent simulation of a public goods game with trust dynamics and reinforcement learning
cola-manipulation-bound research
Analytic upper bound on pool manipulation gain in the COLA draft mechanism
shapley-hri-attribution proposal
Causal credit assignment in human-robot interaction using Shapley values