Gyana Ranjan Padhi profile photo 1 on Tal

Gyana Ranjan Padhi

Distributed database engines | PostgreSQL | Teradata | IDV | Nonstop SQLMX

Greater Bengaluru Area

14+ years of work experience

Uber
Apr 20251 yr 6 mosPresent

Bengaluru, Karnataka, India

Real time analytics platform built over Apache Pinot.
Salesforce
Jun 20222 yrs 11 mosApr 2025

Hyderabad, Telangana, India

Salesforce Database (SDB) is built over PostgreSQL and redesigned with decoupled storage/compute for OLTP to moderate OLAP workloads. I am leading the India query optimizer development efforts for the database engine. Some of my key contribution include monitoring views that provide runtime insights into the state of DB, planner enhancement to limit the failures cause by incompatible plan constraints, predicate pushdown for subqueries, cost based selection of non-folded subplans, rewrite of complex terms and hash join improvements.
Informatica
Mar 20224 mosJun 2022

Hyderabad, Telangana, India

Informatica Data Vault (IDV) serves as a low cost data store for less frequently accessed relational data. The data is stored with high compression ratio in columnar format. IDV allows retrieval of data through its SQL engine without the need for decompressing the entire data set. I designed and implemented multiple enhancements in the query optimizer, metadata engine, memory management, external storage access, and the logging engine. I led the development of Data Vault engine and storage layer through multiple releases along with the customer escalations. Some of my key contributions are # Rainstor migration project: Designed and implemented a lightweight data migration system (migration service, data loader, monitoring service, external scheduler API) for IDV that gave 3X performance throughput. This helped new customers with petabyte scale data migrate to IDV in a realistic time frame. # Cloud Store: Led the development of Azure/AWS as transparent storage options for Data Vault which was vital for moving IDV to cloud platform. # Data Privacy (GDPR) operations: Enhanced IDV to support GDPR mandated operations on stored data in a limited time frame. This helped acquire new customers looking for GDPR compliance. The work involved enhancing the query optimizer and executor to support cross database search, update and delete operations. # Firebird maintenance: Took the initiative of maintaining the source code for open source Firebird DB which served as the metadata store for IDV. This helped save the annual Firebird support license cost. # An ML based query parser for relational and noSQL databases that provided automated recommendations for SQL optimization, syntax/semantic corrections and incremental support for newer syntaxes. # AI chatbot that could translate English language into context sensitive data processing instructions, saving the development time and steep learning time for data transformation platforms.
Informatica
Apr 20202 yrsMar 2022

Greater Hyderabad Area

Informatica
Apr 20182 yrsMar 2020

Hyderabad, Telangana, India

Hewlett Packard Enterprise
Feb 20151 yr 3 mosApr 2016

Greater Bengaluru Area

NonStop SQL/MX is an OLTP database that has a high level of fault tolerance and scalability. This is used in major Stock exchanges, telecom and banking applications that have mission critical execution needs. # ORACLE compatible functions: Implemented a set of functions that were not earlier supported by SQL/MX. This feature will help potential customers, who want to migrate their applications from ORACLE to NonStop SQL/MX. # Predicate simplification: A mechanism for analyzing complex conditions in query. This enhancement intends intelligent reordering of predicates, improving access path selection, reducing condition evaluation overhead and for detecting unsatisfiable conditions that can help in pruning of steps. # A POC to integrate NonStop SQL/MX with Hadoop. Developed a module that transferred NonStop data into Hadoop Cluster for processing and fetch the result, using InfiniBand.
Teradata Corporation
Jul 20104 yrs 7 mosJan 2015

Hyderabad, India

Teradata is one of best Very Large Relational Database Systems (VLRDBMS).I mostly worked on the modules that use the statistics, look-ahead logic, access-path costing and some heuristics to generate the optimal execution plan. Some of my key contributions are to the features 1. Inlist optimization 2. Cost estimation framework. 3. Static and Dynamic partition elimination. 3. Partitioned Primary index. 4. Statistics framework.

Gyana’s AI Stack

claude code
claude code
cursor
cursor
gemini
gemini

Education

B.Tech, Information Technology

biju patnaik university of technology

20064 yrs2010

Join Gyana and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Gyana Ranjan Padhi on Tal

Headline

Distributed database engines | PostgreSQL | Teradata | IDV | Nonstop SQLMX

About

Over the last 16 years, I have mostly worked on designing, implementing, and optimizing distributed database engines, query compilers, and large-scale data processing systems. In my previous 3 roles, I enjoyed the opportunity of starting from scratch and growing with the team learning new things every day. I have spent quite some time on formulating cost/heuristics-based optimization techniques and reverse engineering of complex, inter-dependent code modules to improve performance. Now a days, I am focusing on cloud scale query planning, optimized data access mechanisms and using ML for proactive identification and automated resolution of performance problems. Currently at Uber, I contribute to building a real-time analytics platform leveraging Apache Pinot.

Location

Greater Bengaluru Area

AI stack

claude code

cursor

gemini

Work experience

Staff Software Engineer at Uber, Apr 2025 - Present · 1 yr 6 mos, Bengaluru, Karnataka, India, Real time analytics platform built over Apache Pinot.

Lead Member of Technical Staff at Salesforce, Jun 2022 - Apr 2025 · 2 yrs 11 mos, Hyderabad, Telangana, India, Salesforce Database (SDB) is built over PostgreSQL and redesigned with decoupled storage/compute for OLTP to moderate OLAP workloads. I am leading the India query optimizer development efforts for the database engine. Some of my key contribution include monitoring views that provide runtime insights into the state of DB, planner enhancement to limit the failures cause by incompatible plan constraints, predicate pushdown for subqueries, cost based selection of non-folded subplans, rewrite of complex terms and hash join improvements.

Staff Engineer at Informatica, Mar 2022 - Jun 2022 · 4 mos, Hyderabad, Telangana, India, Informatica Data Vault (IDV) serves as a low cost data store for less frequently accessed relational data. The data is stored with high compression ratio in columnar format. IDV allows retrieval of data through its SQL engine without the need for decompressing the entire data set. I designed and implemented multiple enhancements in the query optimizer, metadata engine, memory management, external storage access, and the logging engine. I led the development of Data Vault engine and storage layer through multiple releases along with the customer escalations. Some of my key contributions are # Rainstor migration project: Designed and implemented a lightweight data migration system (migration service, data loader, monitoring service, external scheduler API) for IDV that gave 3X performance throughput. This helped new customers with petabyte scale data migrate to IDV in a realistic time frame. # Cloud Store: Led the development of Azure/AWS as transparent storage options for Data Vault which was vital for moving IDV to cloud platform. # Data Privacy (GDPR) operations: Enhanced IDV to support GDPR mandated operations on stored data in a limited time frame. This helped acquire new customers looking for GDPR compliance. The work involved enhancing the query optimizer and executor to support cross database search, update and delete operations. # Firebird maintenance: Took the initiative of maintaining the source code for open source Firebird DB which served as the metadata store for IDV. This helped save the annual Firebird support license cost. # An ML based query parser for relational and noSQL databases that provided automated recommendations for SQL optimization, syntax/semantic corrections and incremental support for newer syntaxes. # AI chatbot that could translate English language into context sensitive data processing instructions, saving the development time and steep learning time for data transformation platforms.

Principal Software Engineer at Informatica, Apr 2020 - Mar 2022 · 2 yrs, Greater Hyderabad Area

Lead Software Engineer at Informatica, Apr 2018 - Mar 2020 · 2 yrs, Hyderabad, Telangana, India

Software/Systems Engineer II at Hewlett Packard Enterprise, Feb 2015 - Apr 2016 · 1 yr 3 mos, Greater Bengaluru Area, NonStop SQL/MX is an OLTP database that has a high level of fault tolerance and scalability. This is used in major Stock exchanges, telecom and banking applications that have mission critical execution needs. # ORACLE compatible functions: Implemented a set of functions that were not earlier supported by SQL/MX. This feature will help potential customers, who want to migrate their applications from ORACLE to NonStop SQL/MX. # Predicate simplification: A mechanism for analyzing complex conditions in query. This enhancement intends intelligent reordering of predicates, improving access path selection, reducing condition evaluation overhead and for detecting unsatisfiable conditions that can help in pruning of steps. # A POC to integrate NonStop SQL/MX with Hadoop. Developed a module that transferred NonStop data into Hadoop Cluster for processing and fetch the result, using InfiniBand.

Software Engineer II at Teradata Corporation, Jul 2010 - Jan 2015 · 4 yrs 7 mos, Hyderabad, India, Teradata is one of best Very Large Relational Database Systems (VLRDBMS).I mostly worked on the modules that use the statistics, look-ahead logic, access-path costing and some heuristics to generate the optimal execution plan. Some of my key contributions are to the features 1. Inlist optimization 2. Cost estimation framework. 3. Static and Dynamic partition elimination. 3. Partitioned Primary index. 4. Statistics framework.

Education

B.Tech, Information Technology, biju patnaik university of technology