Arnab Deka

Amazon Web Services (AWS)

Senior SDE (Seattle & Vancouver): Lead Engineer of AWS Chatbot.

Oct 2014 - Present

• Between 2014 and 2016, I led multiple projects in the high-scale critical world of AWS Console Platform.
• AWS Chatbot: Started in 2016 as the first engineer. Working with Principal Engineers and Product Managers I launched a successful product in July'19, while building and growing the team (10+ today).
• In Nov'19 we launched the capability to run AWS commands from Slack.
• Watch this space for more coming soon :)

Nilenso

Senior Technical Architect (Bangalore)

Jul 2014 - Sep 2014

While preparing to move back to the US, I worked with Nilenso to explore their product development ideas, immersed myself in Clojure, and helped flesh out the initial architecture of Kulu.

• Backend: A Swagger based RESTful API built on Clojure, working on top of AWS (S3, SQS) and Elasticsearch and PostgreSQL, capable of running OCR on itemized receipts.
• Frontend: A Rails website to manage user's receipts and visualize data.

LivingSocial

Senior Software Engineer (Bangalore/remote)

Jun 2012 - Jun 2014

As a part of the Internal Apps team, I focused on making the deal production workflow and while improving quality/automation. Effective in a remote role, 10-hours away from most of the team, while still pair-programming!

• Designed and implemented the messaging sub-system and webapp that communicates with AmazonLocal via SQS/S3 and allows users to visually see and manage change requests. Described as "intuitive and flexible" by Product Manager, end-users delighted by UX and the time saved.
• Infrastructure upgrades (ree to Ruby 1.9, Rails 2.3 to Rails 3) on massive Rails applications.

Amazon Web Services (AWS)

Software Development Engineer II (Seattle)

Nov 2010 - May 2012

Built a Ruby/Rails platform for the AWS website and started re-architecting components of the site to use this platform. Influenced the team to use Ruby where appropriate and helped recruiting a great team of engineers.

• Lead the design and development of AWS website's contextual and faceted search component. Combined disjoint datasets in a generic way making new dataset additions easy. Widely appreciated by customers of all sizes and within AWS and got featured in the AWS blog.
• Re-architected the community-oriented components of the AWS website to make them much more customer friendly and usable. Developed a generic catalog framework that is re-usable across multiple areas ( EC2 AMIs , Developer Tools , Sample Code etc.) and started the process of moving the catalogs to the new framework. Drove graphic design and UX discussions with design-teams. Experimented with pair programming and was perceived to be fast and very productive by the team. Code and test's quality was eye-opening for the team.

Amazon.com

Software Development Engineer I & II - Supply Chain (Seattle)

Jan 2008 - Nov 2010

Started my career in Amazon as an SDE I and got promoted within a year. Helped solve complex business problems in supply chain by working with analysts, building software solutions and a varied mix of tools to make sense of the massive amount of data.

• Built a prototype to generate the Amazon.com S&OP forecast with a linear-programming model. Drove adoption by demonstrating that it produced better or in-line results compared to the manual S&OP (which takes 80 hrs/week of analysts).
• Drove adoption of automated-testing and BDD with tools like rspec/cucumber, while being the junior-most engineer. Evangelized other software engineering tools like git, jQuery etc.
• Developed a tool to analyze the amount of wasteful cross-dock happening in Amazon.com fulfillment centers. Designed it applying a binary-search technique to avoid inefficiencies in an API that made the tool 65% faster with no loss of accuracy.
• Developed a RESTful service to get Amazon.com fulfillment network's product-allocation configuration, previously a 2000+ line yaml maintained by SDEs. Also developed a webapp for analysts to make self-service, on-demand configuration changes. Reduced 4 hours/week of SDE operational work.
• Lead the development of a critical reporting webapp that showed alarming trends in terms of labor requirements in Amazon warehouses to cope with the US holiday-season demand.
• Wrote a tool to mine various Amazon data sources to find alternate inventory we could use to fulfill those last-minute "Santa orders". Coped with a very tight deadline and a lot of ambiguity.
• Wrote my first Rails app, in 2008, to hook into a backend system (that prioritizes/schedules inbound trucks into Amazon warehouses) to allow rare but critical manual overrides to the system's scheduling decisions. Estimated to save 120 hours of analyst and SDE time in Q4'2008.
• Wrote an in-house analytics tool to parse apache access logs to help us know more about our customers.

Infosys

Programmer Analyst (Bangalore & NYC)

Jul 2004 - Dec 2007

Started off as a junior programmer in the Goldman Sachs account (in Bangalore and NYC) and went on to build software that caught the customer's eye, leading to major development efforts collaborating with client-side engineers.

• Designed and implemented a system (backend services, batch-processes, webapp) to handle a new SEC regulation for Short Sales. Acknowledged as a flexible and elegant solution to a challenging problem (complicated regulatory rules, real-time interaction with outside systems over EMS messaging). Learnt automated-testing during the project, finished with 93% coverage.
• Re-architected the Canadian and Lat-Am Clearance system with a team of 8 diverse engineers. Lead the overnight batch part of the project. Learnt and practised agile development style.
• Designed and implemented the CNS (DTC) Bulk Settlements component.
• Automated the NSCC Recap process at Goldman - freeing up a day per month of analysts.
• Wrote a tool to diff the team's live DB schema/triggers/stored-procs with source-code definitions. This helped the team avoid occasional, but critical accidents in production, much appreciated by the client.
• Kept Goldman's American Clearance systems running with a team distributed over 3 continents (support/maintenance/enhancements).