FFFF GitHub - nicoloangileri/nicoloangileri: Profile README · GitHub
[go: up one dir, main page]

Skip to content

About

Profile README

Resources

Stars

1 star

Watchers

0 watching

Forks

Latest commit

 

History

15 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 

Repository files navigation

Hi, I'm Nicolò

I build machine-learning and optimization tools for forecasting and quantitative finance

Open source

35 pull requests merged into libraries used across ML, quantum computing and finance

  • OpenAI Tart: container-registry authentication with spaces between WWW-Authenticate directives, as RFC 7235 allows (#1377)
  • NVIDIA cuOpt (2): nightly CI for its CVXPY integration on images without a C++ compiler (#2101), and a minimize default for the objective sense of its MPS data model (#2181)
  • Apple MLX: reductions over very large arrays (32-bit index overflow on CPU and GPU) (#4549)
  • MLX-VLM: save_pretrained on the Muse Glimmer processor, which stopped checkpoint conversion (#2535)
  • Amazon Chronos: series lengths for categorical item ids in predict_df (#537)
  • QuantLib (2): underflow in the non-central chi-square distribution (#2908), and a faster Heston variance mesher for small vol of vol (#2940)
  • CVXPY (7; top external contributor by commits in the 90 days to October 31, 2026, and joint 32nd of 244 contributors by commits all time): silent wrong answers on mixed-integer problems with NVIDIA cuOpt (#3555), MIP warm start for SCIP (#3483), geo_mean on any shape (#3493), tuple axes in std, axis behavior of cvar, axis checks for matrix norms, Gurobi's own error message when it fails before finding a solution (#3574)
  • skfolio (6; 6th of 47 contributors by commits): VaR, EVaR and sample-weight fixes in the risk measures, drawdown contributions, the Bodnar-Okhrin shrinkage of expected returns, and sample weights in the empirical covariance (#417)
  • NautilusTrader (2): wrong default moving averages in trading indicators
  • QuantEcon: ARMA impulse responses and simulations that came out delayed when p > q (#976)
  • Nixtla StatsForecast: finite-sample conformal prediction intervals, shared with the Conformal Seasonal Pool model (#1246)
  • Microsoft FLAML: naive and seasonal naive forecasts that match their definitions (#1611)
  • Unit8 Darts (2): conformal prediction intervals (#3227) and compatibility with pandas 3.1 (#3289)
  • FinRL (AI4Finance): tradability check on the price instead of the first indicator in StockTradingEnv (#1456)
  • Pasqal Pulser: lazy conjugate/negative tensors converted to NumPy without resolving them (#1130)
  • gcsopt (graphs of convex sets): compatibility with the new CVXPY solving chain (#23)
  • Also Qdrant, NumPyro (2) and Qiskit Experiments

In review: NVIDIA cuOpt (MPS writer), AutoGluon, Nixtla StatsForecast, Open GenAI Stack (formerly Llama Stack), SciPy.

Research and projects

Open to ML engineering, quant, software and data roles, including remote, part-time or project-based contracts, and to research projects. LinkedIn

About

Profile README

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

0