Apr 2023 – Oct 2024 · 1 yr 6 mos · Greater Chicago Area · Hybrid

• I continued work on statistical strategies for wide/illiquid market, multicurve trading, market microstructure and regime identification.

Sep 2022 – Apr 2023 · 7 mos · Greater Chicago Area

Sep 2021 – Present · 4 yrs 6 mos · Chicago, Illinois, United States · Hybrid

• I work on image foundation models and world models for histopathology, LLMs for genomic and transcriptomics data and multimodal models combining these. I also investigate applications of quantum computing to biomedical uses of multimodal deep neural models.
• Our earlier work on uncertainty quantification (UQ) of Bayesian Neural Networks led to substantial improvements in the confidence and quality of cancer diagnosis based on image classification and led to a Nature Comm. publication with U.Chicago and Mayo Clinic: "Uncertainty-informed deep learning models enable high-confidence predictions for digital histopathology": https://www.nature.com/articles/s41467-022-34025-x

May 2021 – Sep 2022 · 1 yr 4 mos · United States

• Port22 merged into DV Trading and I'm happy to remain a part of the Port22 teach throughout!
• I work primarily in the futures markets (STIRS, energies) trading in the single microsecond to tens of milliseconds latency range.
• I start with statistical analysis, build and backtests models on scalable cloud clusters, and end with the implementation of the strategy in a highly-optimized low-latency code (kernel bypass, lock-free wherever possible, etc.).
• I have worked on models interfacing with hardware (FPGA) and microwave pipelines, although lately the trend has been towards software-only strategies with more complex pricing and hedging models.
• My latest interests are on strategies based on (multi)curve-based models of market microstructure, particularly with applications to wide and illiquid markets (e.g., TTF Natural Gas).

Dec 2018 – Sep 2022 · 3 yrs 9 mos · Greater Chicago Area

Oct 2017 – Dec 2018 · 1 yr 2 mos · Chicago, Illinois

• Quantitative strategies based on alternative data (Yodlee, Slice: transaction records, credit card receipts, satellite. weather and oil-well data, web scraping data)
• Contributed to the research on a central systematic book
• Developed an NLP analysis system for transaction receipt attribution and regulatory filings' sentiment analysis
• Developed a number of cross-asset systematic strategies, including low-fee passive products

Feb 2017 – Oct 2017 · 8 mos · Chicago, Illinois

• Factor analysis, portfolio construction, transaction cost modeling
• Developed a quant research/portfolio construction library based on the Mosek/cvxopt optimizers

Nov 2015 – Feb 2017 · 1 yr 3 mos · Greater Chicago Area

• Building a scalable distributed trading data ingestion analytics platform in C++, Java, Python, and R. Scalable ETL into Vertica, SciDB, Parquet files. In situ high-performance parallel machine learning in SciDB. Real time stream analytics with Kafka/Kafka Streams.

Dec 2013 – Oct 2015 · 1 yr 10 mos

• Lead a group of postdocs and junior scientists
• Developed mathematical models of materials (condensed and soft matter).
• Carried out large-scale computational analysis of nuclear reactor safety, simulated DNA translocation, dynamics of bacterial suspensions, complex ferroic materials. Employed finite element modeling (FEM), nonlinear constrained optimization, Monte Carlo simulation, statistical analysis of data using feature extraction & tracking and visualization.

Oct 2008 – Oct 2015 · 7 yrs

• Designed and implemented novel numerical algorithms to model complex physical systems. Built HPC libraries for scalable simulation of global climate (ice sheet dynamics). Developed a stochastic model of anomalous viscosity reduction in bacterial suspensions. Proposed a novel nonlinear multiplicative noise model of phase transition in suspensions of microtubules. Shared in an R&D 100 Award for scalable parallel nonlinear solver library PETSc. Supervised graduate students at University of Chicago, Northwestern and Penn State.

Jan 2007 – Dec 2013 · 6 yrs 11 mos

• Researched analyzed and developed scalable nonlinear solvers for simulation of complex physical systems, mostly in materials science, soft matter physics, molecular biology. All code development done in C/C++ with some Fortran and extensive use of Python. Contributed substantial effort to the development of a popular scalable linear/nonlinear solver and optimization PETSc library.

Jan 2005 – Jan 2007 · 2 yrs

• Analyzed instability mechanism behind the circadian cycle in cyanobacteria.
• Designed and implemented highly-accurate algorithms for simulation of the cell energy metabolism, dynamics of interacting biopolymers (microtubules). Developed a library of scalable bifurcation location and tracking algorithms. All algorithms implemented in a fully parallel scalable manner in C, based on MPI and with Matlab & Python bindings.

Dec 2002 – Jan 2005 · 2 yrs 1 mo

• Researched highly accurate algorithms for simulation of plasma dynamics, deep water waves and dynamics of magnetic nanodots. Developed novel time integration methods based on symplectic geometry. All algorithms implemented in a fully-parallel scalable PETSc library.

Jan 2006 – Jan 2011 · 5 yrs

• Proposed and implemented a novel method for calculation of ion densities in cell membrane proteins. The new method had a sharply increased computational accuracy, yet scalable thanks to a novel use of GPU computing.

Jan 1996 – Jan 2002 · 6 yrs

• Researched and designed symplectic and multisymplectic algorithms for time integration of Hamiltonian PDEs. These algorithms have their roots in the deep mathematics behind the so-called "integrable systems", including the soliton PDEs and their symplectic geometry. Laid foundation for a high-performance library of "geometric" time integrators, implemented in C and using MPI.