Aman Thakral Email & Phone Number

Bellevue, Wa, Us

Bellevue, Wa, Us

Leading technical strategy for the payments platform.

Acquired

Bellevue, Wa, Us

Lightsphere AI INC builds advanced Artificial Intelligence and Machine Learning algorithms, products and services for existing businesses and for new consumer and b2b problems, especially in collaboration with data-driven start-ups and medium-size businesses. Many small innovative companies now have a lot of rich data; your data + our algorithms and ML process engineering = immense new dollar value for your business and your customers, and separates you from the pack. Past and current customers include Atigeo, MetaBrite, Womply, and Blue River Technology. Full brochure of products and services we provide and references are available on request. Please contact us at business@lightsphereai.com for further questions.

Johnston, Ia, Us

- Responsible for developing data science infrastructure @ Granular.- Developed a library to help simplify Geospatial data science (https://github.com/granularag/pyspatial). The library was used to build the datasets behind AcreValue.- Developed a tabular data structure (similar to DataFrames in python/R) in JavaScript for developing client side reporting/analytics. The data structure powers much of the reporting functionality at Granular.- Designed a schema driven table UI for interactively working with tabular data. This design now powers many of the tables in the Granular App.- Built a data catalog that serves as a backbone for all of the data-science workflows. Allows users to easily save/load heterogenous datasets that are tagged with rich metadata. Also keeps track of jobs (both ad-hoc and scheduled) executed in our data pipelines.

- Designed and implemented the collaborative machine learning workflow component of Atigeo’s xPatterns platform. Primarily worked on the backend components of processing experiment graphs (in Java), scheduling node executions using Apache Aurora/Mesos, and integrating/automating system deployments (using python/bash). Also made significant contributions to the search and cataloging of data using ElasticSearch.- Researched and developed an online, multi-class algorithm for detecting changes in satelliteimagery using the python scientific stack. The model was exposed via a REST API, andincluded an web-based judgement tool (written in HTML/Javascript/CSS) for collectinglabelled data from human judges. The model server backend was designed to not only behorizontally scalable, but also facilitate model experimentation, measurement, anddeployment.

St. Louis, Missouri, Us

- Researched and implemented the soil hydrology model that is used by all parametric insurance products. This model is used to estimate drought stress, the predominant cause of crop harm in the United States.- Designed and implemented a generalized model parameterization for estimating crop growth stages. This parameterization supports all existing insurance products using a common interface, and provides extensibility to support rapid model development.- Conducted parameter optimization for the crop growth stage models in both the soybeans and winter wheat parametric insurance products. The accuracy of these models is critical as they establish the timing of the weather related crop harm across the midwestern US.- Carried out performance optimization of Clojure based process modeling code to meet latency requirements of policy quoting.- Implemented data flows to process and analyze large scientific datasets using Hadoop/Cascalog.- Developed and implemented a process model to estimate corn harvest moisture as part of Climate.com's decision support tools.

- Developed Europe wide crop disease forecasting models for Bayer CropScience- Developed software for sensor fault detection for WIN's extensive weather network.- Implemented machine learning algorithms for in house data analysis

Hamilton, Ontario, Ca

• Served as a domain expert during the design phase for process modelling, simulation and optimization software• Aided in the design of the relational database• Developed models for testing and debugging

Hamilton, Ontario, Ca

• Administrated Microsoft Office SharePoint Server for the GMC Centre for Engineering Design• Designed a prototype website for an online learning environment, using SharePoint, to be used as an academic portal for chemical engineering (McMaster, Texas A&M, and Qatar University)

Hamilton, Ontario, Ca

• Reviewed course material and evaluated coursework for “Introduction to Reactor Design”• Taught students how to use calculators, Matlab, Excel and Polymath to more efficiently solve assignment problems.

M.Eng Design, Process Systems Design And Operation

B.Eng & Biosci, Chemical Engineering And Bioengineering