We are Kredaro, an early stage startup focussed on solution offering on these following broad areas and their respective sub-categories, our core competency revolves around tailoring solutions for challenges @Scale.
- Building Cloud Native, Scalable, Resilient infrastructure/applications on Kubernetes.
- Systematic performance and resiliency enhancement of complex distributed systems driven by analytics and observability.
- Complete solution starting from streaming/ingesting data from various sources to storing them and analyse for various batch and stream processing needs.
- Customized and scalable solutions to solve value additive and challenging problems using Machine learning and Deep learning.
- Running Composable, Portable and Scalable Machine learning and Deep learning jobs on Cloud/On-Premise (Will be delivering a talk on the same next week at the prestigious Devops days India conference.
- Scale runs through our veins, we value numbers more than gut feeling.
- Minimalism, Simplicity, performance, stability, exhaustive tests and cost effectiveness are the core values which drives our architectural and implementation decisions. These core values lays out a firm foundation for building scalable systems.
- We not just use open source tools to connect the dots, we’ve been part of amazing engineering teams and in collaboration have architected and built cutting edge and disruptive technologies which have been widely used and deployed by millions across the globe.
- Modern and disruptive stack choices which includes Minio, Spark, Tensorflow, Kubernetes, Kubeflow. Python and Go are primary choices at the server side.
- Active participation and leadership in open source communities.
- We are at places where next open source revolution is happening various International forums/conferences, At Google Gophercon India 2016 , At Google Gophercon India 2017 , Building scalable cloud native storage using Minio on Mesosphere container platform .
- Have offered huge cost cuttings on cloud infrastructure by migrating to Kubernetes and building in-house solutions as alternatives for managed solutions.
Few Challenging Use cases:
Our biggest client has been Rapido bike taxi, when our engagement began they had serious infrastructure challenges due to complex architecture containing 50+ micro services. Here were the primary concerns,
- Scalability woes.
- Over the top Cloud bills, Ineffective usage of VMs and managed services were primary reasons.
- No observability in the system.
- Building high performant and cost effective analytics for ever growing data.
- And application of solution has been since it has be done without affecting the currently running business.
Here is the description of few important solutions we built to address the issue,
- Migration to Kubernetes to resolve the scalability issues, it also drastically bought down the bill since Kubernetes offered a consolidated pool of resources which bought effective usage of their systems.
- Built in-house altering, application performance management and tracing capabilities to diagnose the stability and performance issues and use these metrics to attack the bottlenecks in microservices. We used Prometheus, Jaegar, Elastic-APM, Elastic search, Graphing, Kibana along with few custom micro services.
- Log analytics to obtain insights and derive key metrics to take business decisions. We used Fluentd to stream data into Minio (The storage backend) and have used Apache Spark for analysing the data.
- High traffic data ingestion, storage and wide range analytics capabilities for application data. Again Fluentd, Minio and Spark has been at core of our analytical solution offering.
- Building an voice based AI assistant to understand the behaviour and to obtain insights from systems at scale.
- Contributions to Google’s Kubeflow, An opens source Kubernetes native platform for Distributed Machine learning and Deep learning. We are hopeful of its future and placing a bet on the technology.
- Automated alerts and analytics for application server performance.
- Building black box Cost effective and scalable analytics for stream and batch IOT data.
We see a huge market for solutions that could emerge from our expertise. The applicability could range over wide variety of domains and can be used to adddress and solve valuable + challenging problems which could create considerable business impact. If you see a possibility of collaboration, Let’s talk…….. :)
Karthic Rao, Founder CTO, Kredaro