Adam Sanders Email & Phone Number

Equity Trading Team: • Worked with a small team of researchers to launch a new trading group within EG. Developed and maintained research platform as well as production trading processes in three regions. Aided in strategy development, backtesting, data ingestion and research.• Produced​ ​a​ ​distributed​ ​regression​ ​and​ ​analysis​ ​framework​ ​in​ ​Spark​ ​to​ ​load​ ​and​ ​manipulate​ ​data​ ​in-memory​ ​across​ ​a​ ​cluster​ ​of computers​ ​allowing​ ​for​ ​interactive​ ​and​ ​near-instantaneous​ ​intraday​ ​equity​ ​research.• Developed domain-specific language for distributed time-series manipulation and signal definition to aid alpha development.• Created asynchronous distributed model creation, evolution, scoring and deployment system enabling researchers and computers to collaborate on feature definition, model architecture evolution, and hyperparameter exploration.• Developed a variety of new features including forecasts in events, fundamentals, technical analysis, news, sentiment, and other areas. Used linear and non-linear approaches to aggregate hundreds of self-developed features into production forecasts.

Core Technology Team:• First employee and lead developer for the research platform team at new hedge fund. Made extensive use of AWS to create all the technology needed to get up and trading including monitoring, alerts, cron-like processes and long-running services. • Created firm-wide data platform capable of merging terabytes of bitemporal data from multiple sources to produce highly curated security master, pricing, and other market datasets. Pipeline took advantage of Spark, Mesos, and S3 for data management.• Created custom mesos framework for running distributed backtests and other cpu/io intensive tasks. Framework handled dependency management, resource utilization, and code sharing in a clustered environment.

Seattle, Wa, Us

Recommendations Team: • Created the closed-loop optimization pipeline and initial machine-learned models responsible for ranking recommendations at Amazon. Pipeline integrated billions of new training examples into our ML models on a daily basis. Iterated on feature and model development to improve the accuracy of Amazon Recommendations, generating significant incremental revenue in online A/B tests.• Designed and implemented the distributed processing environment used by 40+ developers to easily and flexibly processes the terabytes of data each day required by the algorithms that power personalization features at Amazon using Hadoop, S3, EMR, and other AWS resources. Utilized advancements to create distributed model evaluation framework, as well as distributed implementations of ALS matrix factorization and SGD logistic regression.• Developed a new online recommendations service to power recommendations features, improving flexibility and dropping latencies for key components of our existing service by between 3 and 10x.

Seattle, Wa, Us

Fact Extraction Team:• Worked with a team that specializes in using information retrieval and NLP to extract high quality structured product attributes from billions of merchant contributions with the aim of improving Amazon's global catalog. Developed a distributed clustering application using map reduce to remove duplicates from Amazon's catalog.

Princeton, Nj, Us

Aided in development of AllOurIdeas.org, an online service that allows users to create their own webspace to collaboratively post, evaluate, and rank ideas. Work included using singular value decomposition, clustering, and correlation analysis and collaborative filtering techniques to analyze user data generated at the site.

Princeton, Nj, Us

Founding member and Logistics Manager of 30-member student agency. Oversaw extensive client-designer interaction, supervised agency operations, and developed agency technology and practices.

Princeton, Nj, Us

Developed programs to mine popular interactive websites like eBay, Facebook, and LiveJournal for demographic information. Applied machine learning techniques to locate interesting patterns in these data sets. Advised by Brian Kernighan.

Cambridge, Massachusetts, Us

Worked with large-scale genomic databases to predict the outcomes of genetic transformations with the goal of producing biofuels in bacterial cells. Advised by George Church.

Ann Arbor, Michigan, Us

Developed an artificial intelligence agent to compete in annual Open Source Real-Time Strategy tournament.

A.B., Computer Science

Magna Cum Laude