Rittik Banik

• Working on ZippyDB which is Meta’s persistent, managed key-value storage service, that provides reliable, scalable, geo-replicated storage for high throughput applications. It’s currently deployed in production in 16+ regions, serving out 1.35 billion read/write queries per second and hosting over 267 petabytes of data. It is leveraged across multiple products across Ads, Commerce, FRL, Newsfeed, Instagram, Whatsapp, Messenger etc.
• Lead some of the most challenging migrations for Meta payments including the one that allowed sale of the Meta Quest.
• Landed cross shard transactions for Meta's largest key value store which hosts around 267 PB of data across Meta's family of Apps across Ads, Commerce, FRL, Newsfeed, Instagram, WhatsApp, Messenger etc.
• Lead and directed complex projects at Meta's single largest key value store, ZippyDB including adding improvements to the resharding stack and migration of AI workloads.
• Distributed transactions with 2PC, RAMP and Calvin.
• Working with Rocks DB, LSM storage, bloom filters, skip lists and Compression.

Responsibilities

• Implementation and design of Azure data pipelines to perform Royalties processing based on Azure data processing platforms.
• Influencing and engaging with cross -functional leaders and teams across Engineering, Architecture and Product to define roadmap and segregation of responsibilities.
• Mentoring and providing thought leadership in defining multi year roadmap for multi tenant architecture towards business agility.

Responsibilities:

• Implemented split brain resolution for Apache Ignite using Zookeeper/ZAB.
• Designing and architecting for large scale cloud infrastructure focused on transactional and analytics workloads.
• Event driven transactional workload processing using immutable write ahead logs as primary datastore following polyglot persistence model.
• Performance improvements using lockless data structures which prevent frequent context switches and invalidation of data inside L1 caches.
• Mesh-based mTLS identity, authblue SSO, k8s network policies, and Istio policy management using OPA Gatekeeper.
• Implemented distributed algorithms for message passing systems using asynchronous networks with a focus on appropriate consistency models.
• Debugging and maintaining open source systems which are built on top of distributed algorithms like LWW, Read repair, Linearizable CAS and fault tolerant total order broadcast.

Responsibilities:

• Maintenance and initial development of a strongly eventual consistent system which uses operation based CRDTs and relies on reliable broadcast and commutative operations.
• Debugging complex open source software which used best effort broadcast for state machine replication. This required analyzing complex state update functions which needed to be deterministic, commutative, idempotent and message loss tolerant.
• Design and implementation of a guaranteed message delivery framework replacing existing batch processing with Streaming API's using best effort
• Ensuring durability, correctness, resiliency and high availability allowing for lambda scaling architecture in a cloud native ecosystem.
• Leading a core team of RTF that drove the technology stack for real time processing, in-memory computing, data transformation, advanced analytics and chaos engineering.
• Influencing and engaging with cross-functional leaders and teams across Engineering Architecture and Product to define roadmap and segregation of responsibilities.
• Application of low level Systems programming concepts to provide consultation and direction to the Amex cloud and Information security teams on product evaluation and review.e.x. eBPF.
• Developing custom core components like performant async http dispatchers, JDI debuggers, idempotent checkers and terraform scripts for infrastructure creation.

Responsibilities:

• Leading technology strategy and delivery of a global team working out of UK, US, Singapore and India to standardize Points Transfer experience of MR, Co brand and Cashback cards across 32 markets.

• Implementing Global Loyalty as a Service(GLaaS) using a federated reuse model for market specific integrations.

• Mentoring and providing thought leadership in defining a multiyear roadmap for Redemptions towards business agility and technological innovation.

• Providing data insights into market trends and customer behavior using data driven techniques of Dash-boarding and Predictive Analytics.

• Leading the cloud migration strategy (of J2EE and Golang applications) and implementation to the private enterprise cloud platform (ePaaS) for all loyalty products streams over the last two years.

• Received CIO recognitions for creating Shop Small mobile dashboards using linear regression.

• Spearheaded the effort to introduce Machine Learning into the Loyalty space for automatic anomaly detection and performing market basket analysis.

Responsibilities:

• Prototyping new light weight tools like IBM Liberty Profile and JRebel to automate the delivery process and speed the delivery process during iterations.
• Implementation of Elastic Search Rivers to keep primary and secondary Data sources in sync. Built custom River plugin for IBM Db2 and Elastic Search .
• Implementation of faceted search in Elastic Search with enhanced UX experience.
• Relevance scoring of favorites using function scoring and custom scoring for multiple criteria’s.
• Using decay functions for distance and date based scoring
• Creating dynamic caching strategies using coherence and hibernate 2 nd level cache across clustered environments to enable caching.
  

Responsibilities:

• Data analytics and data mining of telemetry system data sent from millions of GM cars that have the OnStar system installed.
• Creating integrated solutions across the multiple projects (Integrated Logging prototype using Elastic Search, Log back, Play and Scala).
• Code in Groovy to screen request XMLS in order to validate entering requests in XML.
• Utilize dynamic, static routing slips and splitters of Service Mix components on input XML to route them through a series of connecting servers to the backend application. This requires configuring ESB components using context files.
• Architected specific Java 7 security for SSL /TLS handshake for both shared and cert mode communication in a client/server model.
• Used behavioral patterns like Chain of Responsibility, Mediator and Observer across multiple modules to create light weight apps.
• Reengineered structurally outdated apps with adapter, flyweight and decorator patterns.
• Increased concurrency to +60 % using Netty and Java NIO frameworks for the gateways.