I'm the team lead for "HAIL" - HRT AI Labs.We use our deep understanding of statistics, machine learning, optimization and control to broadly improve HRT's trading.We've got real data, challenging problems, clear metrics of success, and lots of computers! If you're interested in joining my team, get in touch.Registered securities trader (SIE+Series 57.)

"Solve intelligence, use it to make the world a better place"Designed and implemented deep reinforcement learning (RL) training systems for a variety of environments, including the Quake 3-based DMLab and StarCraft 2, with a focus on multiagent settings. Research on new training and tuning methods for deep learning, applying RL to combinatorial optimization problems, and applying deep learning to problems in renewable energy and power grids.Tech Lead (June 2017 - July 2018)Team Lead (May 2017 - July 2018)Research Engineer (July 2016 - October 2017)

PhD candidate at the MIT Operations Research Center, supervised by Prof. Dimitris Bertsimas.My thesis work focused on optimization under uncertainty via robust optimization, in particular novel methods for solving problems involving combinatorial decisions across time. Outside my thesis I worked on using machine learning to help solve combinatorial optimization problems, collaborations with industry (including text mining and vehicle routing), and developed software tools to help implement and deploy advanced optimization models.Please see my website http://iaindunning.com for a listing of my research papers, conference talks, software, and technical blog posts.

Teaching assistant for:Spring 2013: Executive MBA class "The Analytics Edge"​Spring 2014: Executive MBA class "Data, Models, and Decisions"​Spring 2015: MBA class "The Analytics Edge"Spring 2016: MBA class "The Analytics Edge"

Worked on search engine, specifically the crawler. Researched, adapted, developed uncertainty-exploiting algorithms to improve the crawl, then analyzed expected impacts using MapReduce over very large (O(10^10) rows, O(10^3) TB) datasets.

Development work on OpenSolver (with Andrew Mason) and Dippy and PuLP (with Michael O'Sullivan).

Teaching AssistantENGSCI 211 - Mathematical Modelling 2 (2011)Tutor for:ENGSCI 233 - Computer Systems and Computational Techniques (2010, 2011)Grading for:ENGGEN 131 - Engineering Computation (2008, 2009, 2010)ENGSCI 211 - Mathematical Modelling 2 (2010)

In this project a computer program (“LPG”) that solves linear programs (a type of optimization problem) with and without the use of these GPUs was written. It was found that if a GPU was utilized, some problems could be solved up to significantly faster than on the CPU. This result suggests that there may be benefits to be had by modifying existing solvers or developing new solvers to use GPUs.

Research in the area of computational acoustics, in particular multiresolution time domain finite difference schemes.

- Developed simulation in Microsoft Excel to simulate workforce training in the NZ primary sector, particularly the dairy industry.- Developed modeling fitting and data exploration tools in Microsoft Excel to fit statistical distributions to data sets of various farming productivity measures.- Worked on an accounting forecast model for an Australasian trading company.

Developed QA system in VBA for inputs and outputs to various industry modelling packages.

Worked on design of Auckland City Council school safety plans, made extensive use of AutoCAD.