Cristian Corrado
Senior Software Engineer
- Engineering better health experiences
- Econometrics tools for central banks
Projects
Econometrics research software
Bayesian Local Projections TBC
Web tool for Bayesian local projections: impulse responses with shape priors, from your own data or FRED series. Built with Prof. James Morley and Naveed Javed.
Beveridge-Nelson Filter
Web tool for the Beveridge-Nelson filter, which splits a time series into trend and cycle. Runs the full model in R behind a Python backend. Built with Prof. James Morley, Prof. Benjamin Wong and Güneş Kamber.
Nowcasting the US Output Gap
Daily nowcasts of the US output gap from a mixed-frequency Bayesian VAR, followed by researchers and central banks. Built with Prof. Benjamin Wong, Prof. James Morley and Prof. Tino Berger.
Algorithms
Best & Second-Best MST
Kruskal's algorithm with a union-by-height, path-compressed disjoint set, and the second-best MST found via lowest common ancestors.
Ukkonen's Suffix Tree
Linear-time suffix tree construction, in both generalised and non-generalised versions.
Boyer-Moore Pattern Matching
Sublinear-time pattern matching, preprocessed with Gusfield's Z-algorithm, plus Hamming-distance approximate matching.
LZ77 Compression
Lempel-Ziv 77 compression with Huffman and Elias omega coding, for 8-bit ASCII files.
Goldbach's Weak Conjecture
Searches for counterexamples using a Miller-Rabin probabilistic primality test built on fast modular exponentiation.
Graph & Dynamic Programming Algorithms
Shallowest spanning tree, constrained shortest paths, a prefix trie and dynamic programming, plus sorting in MIPS assembly.
Systems, tooling & research
jsinspect-plus
Maintained fork of jsinspect, which finds copy-pasted and structurally similar code. Moved to Babel's modern parser with ES2020, TypeScript and TSX support.
Wireless Bushfire Sensor Network
Simulated sensor network that talks to satellites to detect bushfires, built with POSIX threads and OpenMPI.
Colour Dissimilarity Modelling
Data pipeline and Bayesian multidimensional scaling of a large set of colour-similarity judgements, with model selection across metric spaces and dimensions.