Hi, my name is Martin Faust. I connect human judgment and machine intelligence.
At Valsight I build software that helps controlling teams use simulation and AI to make better decisions.
Now
At Valsight we build the foundation modern controlling teams rely on, to get the most out of their data and their instincts. Our software makes the impact of every assumption and measure transparent, all backed by real data and value-driver models.
Research
2012–2016
2012–2016
Before Valsight I researched high-performance in-memory databases at the Hasso Plattner Institute, closely tied to SAP HANA, making very large datasets fast to query and cheap to keep in memory. Our overview paper summarizes the field. My main contributions:
2015
VLDB · IMDM
VLDB · IMDM
Partitioned Bit-Packed Vectors for In-Memory Column Stores
Partitioning bit-packed columns so operations run faster while memory stays tiny.
ACM ↗ 2014
DASFAA · Workshops
DASFAA · Workshops
Vertical Bit-Packing: Optimizing Operations on Bit-Packed Vectors Leveraging SIMD
Arranging bit-packed values so the CPU's SIMD instructions decode and scan them far faster.
SpringerLink ↗ 2014
VLDB · IMDM
VLDB · IMDM
Composite Group-Keys: Space-efficient Indexing of Multiple Columns
A space-efficient way to index several columns together in compressed in-memory column stores.
SpringerLink ↗ 2013
VLDB · IMDM
VLDB · IMDM
Fast Column Scans: Paged Indices for In-Memory Column Stores
Lightweight per-page indices that let queries skip pages that cannot match, speeding up column scans.
SpringerLink ↗ 2012
VLDB · ADMS
VLDB · ADMS
Fast Lookups for In-Memory Column Stores: Group-Key Indices
A compact index for grouping and looking up rows in a compressed column store, cheap to keep up to date.
PDF ↗ Plus co-authored work including Hyrise-NV and Hyrise-R.
All publications on dblp →