Anil Dasari

In the last 7+ years as a solutions architect in HDX Org, I was involved in a wide range of projects, often handling multiple initiatives simultaneously. Some of them are -

CodeGenie: The Gen AI application which started as a helpful assistant that provides code snippets to SFDC developers based on data trained on our proprietary codebase, and quickly evolved into a service leveraged for many other use cases. CodeGenie IDE plugins generate over 2 million events every day, from thousands of daily active users. I designed and implemented the event capture and storage infrastructure to handle this traffic, as well as developed the data processing layer using Python and SQL. This system plays a critical role in tracking CG adoption & acceptance rate, AI utilization at SFDC during SDLC, its impact on developer productivity; driving feature enhancements, CG performance optimizations - like prompt improvements and RAG tuning - and more. I continuously engage with various CG development teams and customers, helping them integrate their services seamlessly with the core CG product.

FIT: I am the PO for ResultNG, which is the backbone for Falcon Integration Testing framework. As part of the recent SCC mandate, I am currently supervising the implementation, rollout and customer adoption of FIT Trust Check process, which is a health check every Hyperforce service must pass in order to exit production deployment moratorium. Besides that, I am working on redesigning FIT's data infrastructure in order to improve its stability, performance and CTS. This project is in design phase.

RDS Initiatives: RDS is the second most expensive billing item on HDX's AWS invoices. In the last two years, as part of an HDX-wide V2MOM measure, I've guided various Hyperforce service teams in significantly reducing their RDS costs, by sizing their instances right, converting to RIs, tuning database parameters and rewriting queries more efficiently. This saved over a million dollars per year. I also helped services build custom DR solutions for their RDS instances (another V2MOM measure).

ABS: From 2016 to Jan 2025, I was the PO for DPX-Data Infra Team, which was responsible for managing over 60 databases and related projects for the Autobuild System (ABS). One of them was LunaDB - an Oracle-backed mammoth of over 60 TB - that's a critical component of ABS, which was migrated from first party to GCP in 2020 with just 8 minutes downtime. To ensure high availability and performance of databases, I continuously engaged with the development teams - guiding them in making the right data organization decisions, developing optimal code, and writing parts of the code if required.

Collaborated with systems, security and network teams to create a reliable, highly available and secure data infrastructure. This included setting up and managing extremely secure database servers that stored highly confidential customer data within a DMZ; setting up two-factor authentication for data access; enforcing SSL authentication; obfuscating all customer data in staging and mirror environments.

Streamlined database releases (configurations, schema changes, application code) by implementing code reviews, enforcing best practices, and integrating industry standard tools like Flyway. This resulted in smooth, successful and timely DB releases.

Planned and executed two major database upgrades (MySQL and Oracle) across multiple data centres with zero downtime. This involved finalizing the target version, testing it out, creating a rollback strategy, coordinating with dependent teams, and rolling out the upgrade to all environments.

Created the Data Infra Team with fresh talent, and groomed the younger members of them into experienced SMTSs. Together, we managed over 125 relational database servers and 4 Hadoop clusters incident-free until I transitioned to a different role.

Setup a 100% redundant DR site for all Data.Com databases in SFDC's east-coast data center. Deployed an Active-Active solution for the Data.Com Enterprise stack.

Devised DR failover strategy and supervised incident-free DR exercises every 6 months. Started with a 15 minute failover window in the first exercise, and reduced it to 5 minutes subsequently.

Managed capacity planning efforts, forecasting future needs and making recommendations to avoid resource constraints.

Worked with a team of external auditors on half-yearly security audits, and contributed to Data.Com achieving SOC2, ISO and PCI certifications.

As the only DBA for Data.Com, helped the TechOps team transition operations from startup mode to enterprise grade, by bringing robustness and resilience to the database infrastructure - which resulted in achieving and maintaining 100% uptime for all production database servers 6 years in a row. This included zero downtime deployments (ZDT).

Provided guidance to one of SFDC partners who was struggling with frequent downtimes - which in turn was affecting Data.Com Clean service - on significantly reducing downtime during releases.

As a senior database engineer at this semiconductor manufacturing firm with a small engineering team that built proprietary CAD software for internal use, my job responsibilities included managing multiple database instances hosted on Linux and Windows-2000 servers, integrating home-built CAD tools with the backend, supervising the build process, and developing Python, PHP and Perl modules.

As a member of Level-3 operations support team for the Siebel based PRM, implemented PartnerPro application. My responsibilities included – provisioning timely hot fixes, system upgrades and enhancements. Most critical responsibility was the successful migration of HP's North American and Latin American partner profiling data from the proprietary Oracle-based Profiler system into PartnerPro.

In a database programming role, handled a reverse-engineering project to replace a legacy application known as California Parent Locator System - developed in the late 1970's for UNISYS/2200 machines using COBOL and JCL– with a robust and efficient Oracle-centered application using PL/SQL.