Apr 2022 – Apr 2025 · 3 yrs · Bengaluru, Karnataka, India · Hybrid

• Working as an architect in the Freddy AI Platform. Working on the ML Platform and ML Ops initiative at Freshworks.
• Driving GenAI adoption within Freshworks to address contextual AI problems, such as summarization and rephrasing, using Large Language Models (LLMs).
• Utilizing LLMs for text classification to respond to user queries and perform tasks conversationally.
• Developing security features: authentication, authorization, and rate limiting.Enhancing scale and performance: designing the platform to support over 1000 queries per second (qps) with low latency, high throughput, and optimized resources.
• Data engineering: capturing significant model metadata and predictions, storing them in cold storage for dashboard generation, accuracy evaluation, and machine learning metrics assessment.
• Establishing a feedback collection framework to gather thumbs up/thumbs down and custom feedback.
• Developing a continuous model monitoring ecosystem for both traditional and GenAI models.

Jun 2019 – Mar 2022 · 2 yrs 9 mos · Bengaluru Area, India

• Domino's enterprise data science platform is used by 20% of the Fortune 100 companies. I am part of the Model Ops engineering which is responsible for the integrated flow of legacy Model Publish/Host product and Relatively new Model Monitor product. I acted both as Tech Lead/Senior Engineer during my tenure in Domino.
• Integrated auth for Model Host and Model Monitor:
• We integrated Model Monitor with RedHat's keycloak ensuring it works seamlessly with Model Host Product. This led to a huge customer impact as customers can access two different products under the same auth.
• Integrated Monitoring Data flow:
• Implemented automatic Data Flow from Domino Model Publish/Host Product to Domino model monitoring flow. This required a complete Batch processing data engineering flow within a very short span of time. Customers could build and monitor models in the same product now. This was the crux of Domino 5.0.0 release.
• Observability:
• Defined the SLA/SLO/SLI for Domino Model Hosting and implemented a pipeline to capture metrics. Fixed latency based performance bottlenecks to improve latency from 10 s to 2 sec helping to retain customers.
• Persistence Layer re-architecture:
• Removed the dependency of external Mongo Connector(Casbah) and Serialization/Deserialization library(salat) and Unified the Data access layer of entire Domino Hosting product. Made the Data access more secure. This has huge impact and received appreciation from top execs.

Sep 2017 – May 2019 · 1 yr 8 mos · Bangalore

• Adobe Storage Solution:
• I was part of this team which was responsible for storage solution at Adobe scale. Owned one micro-service.
• Tech Stack:
• Design principle: Micro-service
• Framework : Spring boot, Spring, Spring Data, swagger
• App server: Tomcat.
• Database: Cassandra.
• Language : Java Enterprise Edition.
• Messaging : AWS SQS, Kafka
• Cloud : AWS.
• Container: Docker.
• Logging/Monitoring: Splunk, NewRelic.
• =============================
• Managed Database Service:
• A multi-cloud approach to solve database as a service. I was part of the cloud infrastructure team to provision and manage Database clusters. Scaled up to MySQL, Cassandra. Was responsible for end to end ownership
• Tech Stack:
• Design principle: Monolith/micro service.
• Framework :Springboot, Spring, ZooKeeper.
• Language: J2EE, Python , ancible, Groovy
• App Server: Tomcat
• Database: MySql/ AWS RDS.
• Messaging : AWS SQS
• Cloud: AWS/Azure

Feb 2015 – Sep 2017 · 2 yrs 7 mos · Hyderabad Area, India

• Bill Payment for Cortana:
• Worked on a scenario to solve Bill payment scenario in Cortana for India Market using Microsoft In-house tools.
• Tech Stack:
• Client: Windows 10, TEE(Text Entity Extraction) Engine.
• Database: Object Store(Document DB in microsoft)
• Backend: XAP(Extendible Application platform).
• Front End: JS, Type Script.
• Tips and Tricks for Bing/Cortana:
• Worked on Tips and Tricks scenario which acts as an advertise platform to send out notifications to Windows HomeScreen, Cortana and other interested Client.
• Tech Stack:
• Database: Object Store(Document DB in microsoft)
• Backend: XAP(Extendible Application platform).

May 2012 – Feb 2015 · 2 yrs 9 mos

• Tech Lead of Samsung Browser framework team. Team Size:5

May 2010 – May 2012 · 2 yrs

• Senior Engineer working in Webkit Open source engine based Android Browser division in smartphone and tablet solutions and feature development

Jul 2008 – Apr 2010 · 1 yr 9 mos

• Worked as a fresher in Regional Feature phone solutions based on efficient algorithms.