I am part of Amazon's Contextual Advertising team responsible for generating Contextual signals (through ML, non-ML techniques), managing their storage and governance. Contextual signals provide context of a Webpage, apps, etc. without the use of user identities(like cookies). I have owned and delivered projects successfully to launch 10+ Contextual signals requiring collaboration and driving alignment on the Problem statement scope and Engineering design along-with SDE-3 and Applied scientists. As part of this, I have designed and build scalable and resilient engineering system from scratch supporting 0.1 Billion lookups per day. I have deployed 3+ ML models handling Production traffic of 1000+ QPS with zero regression/rollback.Additionally, I have launched Data Ingestion API to streamline dataset onboarding to a data-lake and successfully onboarded 7+ Publisher/3P vendor datasets from the legacy ad-hoc flows, enabling data ingestion of 0.5 Billion records per day. As part of the initiative, I have defined the problem scope by diving-deep on the pain-points from previous on-boardings and owned the e2e execution of the SDLC phases.Leveraging my expertise on the legacy systems, I have reviewed 250+ Code reviews and 10+ designs providing impactful feedback focusing on the code maintainability and operational readiness. I have resolved 237 operational tickets (including 141 SEV2s) and drove long term fixes in the legacy system to ensure Production system resiliency. My key strengths lie in Deep-dive, Insist on Highest standards and Bias for action.

Part of Corporate & Investment Banking (CIB) Athena and Trade Risk Services (ATRS) Postrade Team.Flowmonster, the name of my project, is based on automating the trades with clients and the dealers at the same time sending the completed trades to the Regulatory bodies as per rule. It is the trading system that automatically decides when to trade across client and dealer platforms, to provide fast and best-automated price quoting for client quotes, using the liquidity available in the dealer market and in compliance with the current risk.► Reduced client side latency by 30% by removing the redundant API calls through local caching.► Devised the pricing strategy for the client market which provided them prices way lower than the market price at the same time managed to get maximum profit by trading along with the dealers.► Ensured that the risk limit breaches regulated by the regulatory bodies are within the control by devising an algorithm for it.► Communicated really well with team members and had daily scrum calls with them to discuss various issues we, as a team, were facing.► My team reached the finals in the contest, we were runners-up and we had a Profit and Loss (PnL) of $26 M, among two teams who had a positive PnL.Worked on another project called Athena Comparison Tool. The tool provides UI to compare two messages of the same format which is helpful while migrating feeds from one system to another(processes comparing more than 10k deals each day).► Enhanced and improved the Comparison Tool and the report generation so that it could provide more user view.Tech-Stack: Python

Part of Corporate Investment Banking (CIB) Production Management.AaaS (Automation as a Service) Project: The project was to automate the flip from PROD to DR servers so that at any point of time we are sure that the clients aren't blocked waiting for their requests to be served.► Leveraged AaaS micro-services API calls to automate the entire process.► The amount of redundant manual work done by my team was reduced from more than 60 jobs/services to 0 and everything was automated and could be done in a single click. The script managed to flip around 12 application with either Windows or Linux OS.► Every single activity done by my script had been logged so that it could be served as evidence to the regulatory systems.► Ensured that my project could be used not only by my team but also other teams by configuring my code with their application.Tech Stack: Python