Feb 2025 – Jun 2025 · 4 mos

Aug 2022 – Jan 2025 · 2 yrs 5 mos

• Enabled Amazon Q developer capabilities in SageMaker Canvas - a novel way to provide an E2E guided experience for ML practitioners; allowing them to perform data import 40+ data sources, apply 200+ data transformations, allow model training and prediction.
• Leading all the training initiatives within SageMaker Canvas (AWS no-code machine learning interface) as well as in SageMaker Autopilot (AWS low code programmatic automl solution).
• Led the Canvas wide efficiencies to bring the new cost-effective, consistent pricing model for custom model training.
• Launched TimeSeries capablities which is 50% faster and less costly compared to Amazon Forecast. Canvas now also offers model selection for standard build and enables quick and preview build for all users.
• Enhanced SageMaker Canvas's capability by launching Autopilot V2 APIs which is scalable and extendable for future model training use-cases.

Jul 2021 – Aug 2022 · 1 yr 1 mo

• As part of Amazon W/W Returns and ReCommerce team, I continue to expand recommerce inventory platform capabilities to allow automated inventory planning for new businesses based on configurations and events.
• Worked as Amazon's security certifier to verify the security mandates of multiple Amazon's Tier-I systems.

May 2016 – Jun 2021 · 5 yrs 1 mo

• Team: Customer Refund Platform, Ordering Storage, W/W Returns & ReCommerce.
• Build new key-value database for Amazon Customer Refund Platform to store millions refunds placed each day W/W which scales upto millions of reads/writes/updates per second. This allows Amazon to complete the migration of 20+ years of refunds data from legacy systems to this new database.
• Extend the Amazon Ordering data model to incorporate the Refund use-cases which simplifies the execution stack, unified the data modeling and business constructs and helped scaling the system with ever growing data.
• In Ordering Storage team, designed and launched near real-time query engine which performs SQL query on unstructured data and provide results within seconds. This involves live indexing on ever growing Amazon Ordering data, mapping SQL query to ordering data model and processing large unstructured data within seconds.
• Was one of the core contributor of building entire ReCommerce Platform which aims to track E2E inventory lifecycle from Customer returns to injecting back to Amazon Ordering pipeline. This enables Amazon FBA Grade & ReSell, BStock and other business initiative to track their inventory defects and waste to optimize their operations and help Amazon in global sustainability initiative.

Feb 2016 – May 2016 · 3 mos · Noida Area, India · On-site

• Working as part of Acrobat team, responsible for delivering the best document solutions.
• Roles & Responsibilities:-
• 1. Working with product management team and UX team to design new features.
• 2. Mentor & lead team members to help them grow technically & professionally.
• 3. Doing research and design new features to enhance the user experience.
• 5. Working with customers to get feedback & to resolve their concerns
• Major Projects:
• Developed 64 bit PDF Shell to enable thumbnail previews of PDFs on 64 bit Windows machines.

Dec 2012 – Jan 2016 · 3 yrs 1 mo · Noida Area, India · On-site

• 1. Developed Firefox add-on for converting web page to PDF as per new multiprocess architecture.
• 2. Automated the new Chrome "Create to PDF" extension.
• 3. Implemented the Combine App with brand new UI.
• 4. Developed one of the top featured app "Organize Pages" in the latest release of Acrobat : Acrobat Document Cloud.
• 5. Fully automatic updates for Acrobat/Reader on MAC using Adobe Refresh Manager (updater).
• 6. Recent Files Everywhere : Syncing recent files across all the connected devices
• 7. Designing & Developing better Page 0 handling for PDFs.
• Besides that, worked on numerous components of Acrobat - handling Acrobat Headlights data capturing, maintaining and fixing security issues in 3D feature. Also, maintained legacy codebase of second most used feature of Acrobat - Printing and now mentoring others in this area.

Jun 2012 – Nov 2012 · 5 mos · Noida Area, India

• Worked as a part Android E-Mail application development team. Developed ThreePane Layout for email application.

Jun 2011 – Aug 2011 · 2 mos · Montreal, Canada Area

• Computer simulated speckle pattern generation and Determine the parameters for speckle pattern generation.
• Worked with Marin Lagace, research, Hydro-Quebec Research Institute (IREQ) to develop his own Image Analysis software RML-IMAGE (www.rml-image.com). Have Written an additional code for this software to generate differnet Speckle Pattern.

Jun 2008 – Apr 2012 · 3 yrs 10 mos · Jadavpur, Kolkata, India

• 4th year Undergraduate Student,
• Department of Computer Science And Engineering.