Gokul Ganesan
Senior Backend Engineer | Building and Operating Production Geospatial Pipelines | Spatial Observability
Bengaluru
9+ years of work experience
Nov 20241 yr 11 mosPresent
Bengaluru, Karnataka, India
Senior Geospatial Data Engineer - Value Chain Carbon Removal Systems
• Building value chain carbon removal estimation pipelines using ROTHC carbon sequestration models and Dead Organic Matter decomposition analysis across corporate agricultural value chains
• Architecting cloud-native MRV infrastructure on AWS for automated monitoring, reporting, and verification of carbon removals throughout enterprise value chain operations
• Engineering Zarr-based data processing systems for multi-dimensional environmental datasets enabling real-time carbon removal quantification across distributed value chain locations
• Developing end-to-end sustainability analytics platform integrating satellite imagery, soil data, and climate models to deliver value chain carbon removal insights for corporate environmental reporting
Oct 20222 yrs 1 moOct 2024
India
- Engineered high-scale data ingestion infrastructure processing 3M+ developer profiles daily using TypeScript and Google Pub/Sub, achieving 84x performance improvement (14 hours → 10 minutes)
- Architected unified data platform consolidating mission-critical datasets from multiple sources (spreadsheets, databases, SalesForce) into single source of truth dashboard
- Built enterprise dashboard using TypeScript (NestJs), MySQL, BigQuery integration deployed on GKE, reducing client conflict resolution from weeks to days
- Delivered 10%+ gross revenue recovery through automated data standardization and dispute prevention systems
- Orchestrated weekly release cycles enabling rapid iteration and feedback loops for data platform improvements
Jan 20202 yrs 10 mosOct 2022
Greater Bengaluru Area
- Built Imagery warehouse to extract, transform and load 20 year corpus of billions of aerial images in the order of PetaBytes
- Used micro-services architecture to make the ingestion and extraction of the images seamless while reducing the cost per image fetch
- Developing RESTful/gRPC API endpoints to communicate with customers (REST)and between the services (gRPC)
- Working towards reducing the storage cost of the big corpus of imagery data by making the storage as efficient as possible in AWS S3 using better compression.
- Developed services which can transform the raw pixels into a geospatially accurate dataset on the fly within 50 ms per tile.
Jan 20191 yr 1 moJan 2020
Greater Bengaluru Area
- Building a Data processing pipeline in AWS which is able to process terabytes of satellite data efficiently and reduce the data preparation time by 75%
- Deploying an automatic scheduler using Apache Airflow which will schedule jobs based on a predefined schedule reducing the manual effort to run the jobs by 90%
- Setting up a Restful API using AWS lambda and API-gateway to deliver the AI & ML model outputs at scale to the end user with the least amount of human platform monitoring/maintenance possible and as an added benefit this reduced the cost of running the API by 99% (traffic is less than 1M requests per month)
- Building a data rendering engine to render both vector and raster data on to web using tileserver-gl, tippecanoe, GDAL and Rasterio
Jul 20171 yr 6 mosDec 2018
Bengaluru Area, India
- Built a Data specification/workflow customised for mapbox search to normalise all the external vendor data before benchmarking it for quality
- Deployed the data updates and repairs every week after quality benchmarking into multi region production stacks serving close to 1 billion requests a day without any service interruption
- Built a vandalism prevention pipeline for crowd sourced data to prevent malformed or abusive data from entering into the production servers
- Developed a data benchmarking pipeline for different kinds of vector data from point addresses to zipcodes to places to states to countries which helped mapbox to develop a data model which is customised and optimised specifically for mapbox.
- Improved the data testing environment to get better insights about data coverage based on both data quantity and quality.
- Built a robust pipeline powered by AWS lambdas and AWS kinesis stream to handle the constant inflow of data from the community driven crowd sourced endpoints like wikidata, OSM, OpenAddresses etc
- Built a backend data pipeline which can Ingest 400M data points and do a series of transformations (including cleaning, scoring, common data model conversion, QA etc) reliably at scale using various AWS cloud based infrastructure within a couple of hours.
- Maintained all the mission critical infrastructures which handled the loading, transforming and serving the data to mapbox search product.
Nov 20169 mosJul 2017
Bengaluru, Karnataka, India
- Improved and maintained the underlying spatial data which powers all Mapbox products.
- Handled the user feedback towards possible map inaccuracies including but not limited to missing features, incorrect tags and vandalisms using a machine learning model to filter out 95% of the all spam
- Improving and repairing base-data layer without violating the service-level-agreement .
- Created custom filters to filter out all the noisy unwanted data before it reaches the production data foundation and in turn prevent them from rendered on to the maps
Gokul’s AI Stack
Education
Master of Technology (M.Tech.), Remote Sensing and GIS
Indian Institute of Remote Sensing
20142 yrs2016
Bachelor of Technology (B.Tech.), Geology/Earth Science, General
Bharathidasan University
20104 yrs2014
Join Gokul and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal Boss