# Rohith Reddy Kota

> Rohith Reddy Kota is a Boston-based Forward Deployed Engineer and Senior Data Engineer at Rill Data, and one of its earliest engineers. He manages 2+ petabytes of data on Kubernetes across ClickHouse, Apache Druid, and DuckDB, and deploys, migrates, upgrades, and supports real-time and batch analytics for enterprise AdTech and marketing-analytics companies.

Rohith Reddy Kota (also known as Rohith Kumar Reddy Kota or Rohith Kota) works as a Senior Data Engineer, Forward Deployed & Solutions Engineering at Rill Data, Inc. (https://www.rilldata.com), based in Boston, MA. He joined Rill Data in 2021 as one of its earliest engineers and has 9+ years in data platforms and infrastructure, the last 5+ forward-deployed with enterprise customers.

## Key facts

- 9+ years in data platforms and infrastructure (since 2016, across four companies)
- 5+ years forward-deployed with customers (at Rill Data, since 2021)
- 50+ enterprise customers taken live (from scoping to production and support)
- 150+ proofs of concept with customers (run with Sales and GTM)
- 2+ PB on Kubernetes (ClickHouse, Druid, and DuckDB on GKE)
- 30+ ClickHouse & Druid clusters (17 Druid and 14 ClickHouse, sharded and replicated)
- 30–40% less storage & CPU (from schema, compaction, and right-sizing)
- <30s to scan TBs of data, with database optimizations (TB-scale scans after sort keys, projections, and codecs)
- Current role: Senior Data Engineer, Forward Deployed & Solutions Engineering, Rill Data, Inc., since April 2021
- Location: Boston, MA, United States
- Industries: AdTech and marketing analytics
- Languages: English, Hindi, Telugu, German (elementary)
- Website: https://rohithreddykota.com

## Pages

- [Home](https://rohithreddykota.com/): overview, customers, capabilities, product and customer success, how he operates, experience, projects, writing, and contact
- [About](https://rohithreddykota.com/about/): long-form description of his work and the full FAQ
- [Projects](https://rohithreddykota.com/projects/): 11 case studies with problem, solution, impact, architecture, and skills
- [Writing](https://rohithreddykota.com/blog/): technical deep dives, also available as [RSS](https://rohithreddykota.com/rss.xml)
- [Résumé](https://rohithreddykota.com/resume/), also as a [PDF](https://rohithreddykota.com/rohith-reddy-kota-resume.pdf)
- [Full text for LLMs](https://rohithreddykota.com/llms-full.txt): every section, case study, answer, and post in full

## What he does

### Enterprise deployments

Scoping to go-live for 50+ enterprise customers, after 150+ POCs.

- Customer GitOps repos, dedicated clusters, and BYOC installs
- Connectors, pipelines, models, dashboards, alerts, and APIs
- Pre-sales POCs, demos, and solution design with Sales and GTM

### SSO, identity & security

Each user sees only the data they should.

- SSO over SAML, OIDC, and OAuth2
- Row-level security and dashboard access policies
- Least-privilege S3 / GCS access with IAM and service accounts

### Migrations & reverse ETL

Moving data in both directions, with no data loss.

- Legacy systems → the new platform, with planned cutovers
- Snowflake, Redshift, Iceberg, Delta Lake → ClickHouse / Druid
- Reverse ETL back to warehouses and object storage

### Upgrades at every layer

Production analytics keep running during upgrades.

- Kubernetes (GKE), node pools, operators, and Helm charts
- ClickHouse and Druid binary upgrades across sharded clusters
- Rill runtime and customer project upgrades, staged via GitOps

### Performance & cost

~30–40% less storage and CPU, and TB-scale scans in under 30 seconds.

- Sort keys that prune granules, plus partitions, projections, and codecs
- Merges, compaction, and TTLs tuned so part counts stay healthy
- Right-sized node pools and a usage-billing pipeline for cost visibility

### Pipelines & observability

Batch and real-time ingestion at GBs per hour.

- Kafka, Beam on Dataflow, Flink, Airflow, and dbt
- Idempotent incremental loads, backfills, and failure recovery
- Fleet dashboards, SLOs, burn-rate alerts, and runbooks

### AI & developer tooling

Tooling that makes each deployment faster than the last.

- MCP servers and agent skills for Claude Code and Cursor
- Production AI on-call triage service (FastAPI + LLM structured outputs)
- Data-integrity checks that reconcile metrics across pipeline stages

### Petabyte scale on Kubernetes

2+ PB across ClickHouse, Druid, and DuckDB.

- 14 sharded, replicated ClickHouse clusters on Keeper and 17 Druid clusters, all on GKE
- Helm, Terraform, and custom Go operators for cluster lifecycle
- Pod, node, and operator faults debugged in multi-tenant clusters

## Product management, customer success, and proofs of concept

### Proofs of concept & pre-sales (150+ POCs with customers)

Proves the product works on a prospect’s real problem before they commit.

- Partner with Sales and GTM on discovery, solution design, and technical demos.
- Scope each POC with the customer’s technical and executive stakeholders, with success criteria agreed up front.
- Stand up the prospect’s data sources, models, and dashboards on Rill, and answer their security and architecture questions.
- Load-test and benchmark against realistic dashboard traffic to size clusters before go-live.
- Carry winning POCs forward into production deployments.

### Product management (50+ accounts feeding the roadmap)

Brings what customers need back into the product, and helps ship it.

- Gather requirements in recurring account reviews and turn them into roadmap input for Product and Engineering.
- Prioritize and ship customer-driven integrations beyond the core product: Snowflake, Redshift, Iceberg, Delta Lake, S3, GCS, and Kafka sources, plus reverse ETL.
- Designed incremental ingestion for Rill’s ClickHouse driver from customer requirements, the foundation for ClickHouse data modeling in Rill.
- Write and review design docs, and break features into epics, stories, and estimates.
- Turn repeated field work into product and tooling: MCP servers, agent skills, deployment templates, and a data-alert generator.

### Customer success (50+ enterprise customers onboarded)

Stays the technical owner of each account long after go-live.

- Onboard new organizations and users: access, SSO, roles, and first projects.
- Train engineers, analysts, and business users to build and change their own projects and dashboards.
- Run recurring account calls covering the roadmap, new-feature demos, and requirements.
- Act as the escalation point: reproduce, trace across dashboards, queries, pipelines, and infrastructure, fix, then follow up with docs or training.
- Plan migrations, upgrades, and maintenance windows with each customer, and confirm everything works afterwards.
- Give customers usage transparency through the usage-billing pipeline and usage dashboards, and agent-ready repos so they can change dashboards with Claude Code or Cursor.

## How he operates

### Has built at an early-stage startup, without a playbook

He joined Rill Data in 2021 as one of its earliest engineers. Nobody had written the playbook for running petabyte-scale real-time analytics for enterprise customers, so he built the infrastructure, the processes, and the playbook as the company grew.

Evidence:
- Built core platform infrastructure from scratch while stabilizing inherited Kafka, Beam, and Druid pipelines.
- Grew the deployment practice from the first customers to 50+ enterprise customers on 30+ clusters.
- Owned US time-zone engineering coverage on a distributed team.

### Five years as a forward deployed and solutions engineer

At Rill he sits between the customer, Sales, Product, and Engineering. Rill’s customers include The Trade Desk, Comcast, AT&T, AppLovin, and FreeWheel, demanding AdTech and media teams whose dashboards drive revenue every day.

Evidence:
- 150+ proofs of concept and pre-sales engagements with Sales and GTM.
- 50+ enterprise customers taken from scoping to production and supported after go-live.
- SSO, security reviews, migrations, and executive account reviews owned end to end.
- See: https://rohithreddykota.com/#customer-work

### Fixes broken systems and processes without being asked

When a gap costs the team time or money, he doesn’t wait for a ticket. He measures it, fixes it, and makes the fix stick with process or tooling.

Evidence:
- Introduced SLOs, burn-rate alerts, runbooks, and postmortems for the whole database fleet.
- Started cross-cluster storage and cost audits that cut storage and CPU by ~30–40%.
- Built an AI on-call triage agent when alert noise was eating on-call time.
- Replaced hand-built pipeline alerts with a template-driven alert generator.
- See: https://rohithreddykota.com/projects/ai-on-call-triage-agent/

### Finds structure by doing, not by waiting

Most of his work starts with an unfamiliar customer stack and an unclear problem. He gets hands-on fast, builds a working version, and lets the structure emerge from what he learns.

Evidence:
- Made production and customer decisions alone during US hours on a distributed team.
- Every POC starts from a new customer’s data, stack, and constraints.
- Built a BigQuery-to-ClickHouse HLL sketch converter from both engines’ source code, where no documented method existed.
- See: https://rohithreddykota.com/blog/bigquery-hll-sketches-to-clickhouse-uniqcombined64/

### Bias to action: identifies the gap, fixes it, and documents it

A fix isn’t done until the next person doesn’t need him. Every change he ships comes with the runbook, README, or guide that explains it.

Evidence:
- Runbooks for upgrades, replica recovery, scaling, and ingestion failures.
- READMEs and AGENTS.md / CLAUDE.md files in every customer repository, so customers can self-serve with AI tools.
- Troubleshooting guides that start from the symptom, not the code.

### Builds his own projects and keeps iterating on them

Outside his day job he builds, publishes, and keeps improving things, from a tutorial site for developers to hackathon products and open-source tools.

Evidence:
- whiletrue.live: a site of blogs, articles, and tutorials for software engineers.
- rill-agent-skills: installable agent skills for Claude Code and Cursor, iterated as Rill evolves.
- Capital47: cross-border payments with smart contracts and Capital One APIs.
- ComplyTech: blockchain rewards for proper medical-waste disposal, built on VeChainThor.
- See: https://rohithreddykota.com/projects/

## How he runs an enterprise engagement

Every one of his 50+ enterprise customers, and the 150+ proofs of concept before them, followed the same seven steps:

1. **Prove:** A proof of concept on the customer’s real problem, with success criteria agreed up front.
2. **Scope:** Requirements with technical and executive stakeholders, turned into a data architecture.
3. **Secure:** SSO, row-level security, and least-privilege access to cloud storage.
4. **Migrate:** Connect or move data from legacy systems, warehouses, and lakehouses.
5. **Build:** Batch and real-time ingestion, models in ClickHouse or Druid, and dashboards.
6. **Ship:** GitOps-managed dedicated cluster or a BYOC install, taken live.
7. **Support:** Train users, hand off to steady state, and stay the technical owner of the account.

## Experience

### Senior Data Engineer, Forward Deployed & Solutions Engineering, Rill Data, Inc. (2021–present)

Boston, MA · Remote (US). One of the earliest engineers at Rill, an open-source, real-time operational BI platform. He owns forward-deployed engineering and petabyte-scale data infrastructure for enterprise AdTech and marketing-analytics customers, and owned US time-zone engineering coverage on a distributed team.

**Customer engineering**
- Led deployments for 50+ enterprise customers from scoping to go-live and handoff, through customer GitOps repos, dedicated clusters, and BYOC installs.
- Ran 150+ proofs of concept with prospective and existing customers alongside Sales and GTM, from solution design and demos to a working deployment.
- Set up SSO (SAML / OIDC / OAuth2), row-level security, and S3 / GCS IAM, and resolve security and access issues.
- Migrate customers from legacy systems to the new platform, and move data between Snowflake or lakehouses and ClickHouse or Druid, including reverse ETL.

**Product & customer success**
- Turn requirements from 50+ enterprise accounts into roadmap input, and prioritize customer-driven integrations with Product and Engineering.
- Designed idempotent incremental ingestion for Rill’s ClickHouse driver, a feature that came straight from customer requirements.
- Primary technical contact for enterprise accounts: onboarding, user training, executive account reviews, feature demos, and escalations.
- Write and review design docs, and break work into epics, stories, and estimates.

**Data infrastructure**
- Manage 2+ PB on Kubernetes across ClickHouse, Druid, and DuckDB, including 30+ clusters (17 Druid, 14 ClickHouse) on GKE with Helm, Terraform, and custom Go operators.
- Cut storage and CPU by ~30–40% for enterprise customers through ingestion, compaction, and schema changes.
- Built fleet observability, SLOs, the usage-billing pipeline, MCP/AI tooling, and an AI on-call bot. Player-coach to 2 engineers.

Technologies: ClickHouse, Apache Druid, Kubernetes, Go, GitOps, SSO, Kafka, Beam

### Data Engineer, Saltside Technologies (2019–2021)

Bengaluru, India. Streaming, warehousing, and ML for a classifieds marketplace.

- Built an ML-based ad-moderation service that automatically rejected fraudulent listings, reducing manual moderation work.
- Built stateful Apache Flink streaming over real-time web and mobile clickstream.
- Owned Airflow ETL/ELT, the AWS Redshift warehouse, and Tableau reporting for marketing teams.
- Designed APIs between the data platform and product backends.

Technologies: Apache Flink, Airflow, Redshift, Tableau, ML

### Software Development Engineer, Nanoprecise Data Services (2018–2019)

Bengaluru, India. Data-intensive analytics over industrial IoT sensor data.

- Built streaming and data-intensive analytics in Java and Python.
- Ran multi-node Hadoop and Spark clusters and migrated the data stack to AWS EMR.
- Built Go services on MongoDB and gRPC integrations with product teams.

Technologies: Spark, Hadoop, AWS EMR, Go, gRPC

### Technical Associate, Genpact (2016–2018)

Hyderabad, India. Where his data-migration work started.

- Migrated ERP data into SAP R/3 using Informatica, the BackOffice Associates Data Stewardship Platform, and MS SQL Server for staging and batch processing.

Technologies: Data Migration, Informatica, SQL Server

## Case studies

Index: https://rohithreddykota.com/projects/

### Self-hosted ClickHouse platform on Kubernetes

https://rohithreddykota.com/projects/self-hosted-clickhouse-platform-kubernetes/ · Platform · 14 production ClickHouse clusters

A self-hosted, multi-tenant ClickHouse platform on Google Kubernetes Engine that runs 14 production ClickHouse clusters for enterprise real-time analytics, as part of a 2+ PB OLAP fleet.

**Impact:** Dedicated ClickHouse clusters for customers, delivered through a repeatable GitOps workflow instead of hand-built infrastructure, with TB-scale scans in under 30 seconds and ~30–40% lower storage and CPU after tuning.

**Architecture:** GitOps repo → Helm + Terraform → ClickHouse operator → Sharded ClickHouse
**Skills:** ClickHouse, Kubernetes, Helm, GitOps, Distributed Databases
**Stack:** ClickHouse, ClickHouse Keeper, Kubernetes (GKE), ClickHouse operator, Helm, Terraform, GitHub Actions, Go

### AI-assisted ClickHouse optimization with MCP

https://rohithreddykota.com/projects/ai-clickhouse-optimization-mcp/ · AI · Multi-TB single columns found in one audit pass

An MCP-based AI toolkit that lets Claude Code audit and optimize ClickHouse clusters across the fleet, finding the columns, tables, and queries that drive storage and compute cost and tracing each one back to the data model that created it.

**Impact:** A multi-day manual review became a repeatable, agent-driven audit. One pass surfaced single columns holding multiple terabytes, the starting point for cost-cutting work.

**Architecture:** System tables → MCP server → Claude Code + skills → Remediation plan
**Skills:** Model Context Protocol (MCP), ClickHouse, Database Performance Tuning, Cost Optimization, AI Agents
**Stack:** MCP, Claude Code, ClickHouse system tables, Python, SQL, Rill

### Legacy analytics platform migration to Rill

https://rohithreddykota.com/projects/legacy-analytics-migration-to-rill/ · Migration · 1:1 metric parity with the legacy system

Led migrations of enterprise customers from a legacy analytics platform and older pipelines to the new Rill application, moving data, metrics, and dashboards without disrupting daily reporting.

**Impact:** Customers moved to a modern, code-defined analytics stack with numbers matching the old system, faster dashboards, and self-service editing, with the same technical owner through onboarding and support.

**Architecture:** Legacy platform → Parity queries → Rill models + dashboards → Trained users
**Skills:** Legacy System Migration, Data Migration, Apache Druid, Customer Onboarding, Stakeholder Management
**Stack:** Rill, Apache Druid, ClickHouse, DuckDB, SQL, YAML, GitHub, SSO (SAML/OIDC)

### Real-time AdTech bidstream pipelines

https://rohithreddykota.com/projects/real-time-adtech-bidstream-pipelines/ · Streaming · GBs/hr of AdTech events into real-time dashboards

Streaming and batch pipelines that ingest high-volume programmatic-advertising data (bid requests, wins, impressions, and auctions) into Apache Druid and ClickHouse for real-time analytics.

**Impact:** GBs per hour of AdTech events land in real-time dashboards, with a clear path from any dashboard number back to the raw field that produced it. The inherited legacy pipelines were stabilized and their deployment modernized along the way.

**Architecture:** Scala intake → Kafka → Beam on Dataflow → Druid / ClickHouse
**Skills:** Apache Kafka, Apache Beam, Google Cloud Dataflow, Stream Processing, Apache Druid
**Stack:** Scala, Apache Beam / Scio, Google Cloud Dataflow, Apache Kafka, Apache Airflow, Apache Druid, ClickHouse, AWS S3, GCS, Kubernetes

### Warehouse & lakehouse to ClickHouse, with reverse ETL and integrity checks

https://rohithreddykota.com/projects/warehouse-lakehouse-to-clickhouse-reverse-etl/ · Migration · Every hop reconciled for drift, hourly

Migration paths that move analytics data from Snowflake and lakehouse tables into ClickHouse and Apache Druid, and back out through reverse ETL, with automated checks that confirm the numbers match at every hop.

**Impact:** Warehouse-grade data became real-time analytics at a fraction of the query cost, and every stage of the pipeline is continuously checked for drift.

**Architecture:** Snowflake / lakehouse → S3 / GCS → ClickHouse → Reverse ETL
**Skills:** Snowflake, ClickHouse, Reverse ETL, Data Lakehouse, Data Quality
**Stack:** Snowflake, Apache Iceberg, Delta Lake, Parquet, AWS S3, GCS, ClickHouse, Apache Druid, DuckDB, SQL, Rill

### ClickHouse performance & cost re-architecture with projections

https://rohithreddykota.com/projects/clickhouse-projections-performance-cost/ · Performance · 0 dashboard out-of-memory failures after rollout

Re-architected a high-volume ClickHouse analytics workload around tiered projections, directory-based partitions, and tuned compression, which stopped out-of-memory dashboard failures and cut storage and compute.

**Impact:** Dashboard memory failures stopped and queries moved onto small pre-aggregated projections. This work is part of the ~30–40% storage and CPU reduction delivered for enterprise customers.

**Architecture:** query_log profile → Tiered projections → Partitions + ZSTD + TTL → Fast dashboards
**Skills:** ClickHouse, Query Optimization, Schema Design, Cost Optimization, Data Modeling
**Stack:** ClickHouse projections, MergeTree, TTL, ZSTD, SQL, Rill metrics views, query_log analysis

### AI on-call triage agent

https://rohithreddykota.com/projects/ai-on-call-triage-agent/ · AI · Auto-triage alerts classified, deduplicated, and ticketed

A production AI on-call agent that turns monitoring alerts into triaged, deduplicated tickets and Slack updates, so engineers start from a diagnosis instead of a raw alert.

**Impact:** Less manual triage, fewer duplicate tickets, and faster time to a first diagnosis for platform and customer-project failures.

**Architecture:** Monitoring webhook → FastAPI → LLM structured output → Ticket + Slack
**Skills:** LLM Integration, FastAPI, Incident Management, AI Agents, Python
**Stack:** Python, FastAPI, LLM APIs (structured outputs), Datadog webhooks, Linear, Slack, Docker, Kubernetes, Helm

### Fleet observability, data-freshness alerting & usage billing

https://rohithreddykota.com/projects/fleet-observability-freshness-usage-billing/ · Observability · 30+ clusters monitored from one values file

The observability layer for a 30+ cluster, 2+ PB ClickHouse and Druid fleet: templated monitoring dashboards, data-freshness alerting, SLOs, and a usage-billing pipeline that shows customers exactly what they consume.

**Impact:** Issues get caught before customers see them, adding a new cluster or pipeline to monitoring is a config change, and cost and usage are visible to both the business and customers.

**Architecture:** values.yaml → Jinja templates → Rill dashboards → Alerts + billing
**Skills:** Observability, Site Reliability Engineering, SLOs, ClickHouse, Apache Druid
**Stack:** Rill, ClickHouse, Apache Druid, Jinja2, Python, GitHub Actions, YAML, SQL

### Incremental ingestion for Rill’s ClickHouse driver

https://rohithreddykota.com/projects/incremental-ingestion-rill-clickhouse-driver/ · Platform · Open source shipped in rilldata/rill

Designed the incremental ingestion strategy for the ClickHouse driver in Rill, an open-source real-time BI platform, which became the foundation for ClickHouse data modeling in Rill.

**Impact:** ClickHouse models in Rill refresh incrementally and safely, and the approach is available to every Rill user through the open-source project.

**Architecture:** Partition-aware loads → Idempotent models → Rill ClickHouse driver
**Skills:** ClickHouse, Go, Open Source, Data Engineering, Distributed Systems
**Stack:** Go, ClickHouse, Rill, SQL
**Source:** https://github.com/rilldata/rill

### Enterprise deployment blueprint: SSO, row-level security & agent-ready repos

https://rohithreddykota.com/projects/enterprise-deployment-blueprint-sso-rls/ · Platform · Days, not weeks to onboard an enterprise customer

A reusable blueprint for onboarding enterprise customers to Rill in days instead of weeks: SSO, role-based access and row-level security, scoped cloud-storage access, GitOps project templates, and agent-ready repos.

**Impact:** Enterprise onboarding became a repeatable blueprint instead of a bespoke project, and customers can safely change their own dashboards with AI tools.

**Architecture:** SSO (SAML / OIDC) → Access policies + RLS → GitOps templates → Agent-ready repo
**Skills:** Single Sign-On (SSO), Row-Level Security, GitOps, Customer Onboarding, Technical Documentation
**Stack:** SAML, OIDC, OAuth2, Rill, GitHub, AWS S3, GCS, Claude Code, Cursor

### Distributed load testing & performance benchmarking

https://rohithreddykota.com/projects/distributed-load-testing-benchmarking/ · Performance · Before go-live every engine, schema, and cluster size tested

A containerized k6 load-testing harness on Kubernetes that replays realistic dashboard query traffic against Rill and its OLAP backends, plus benchmarks for comparing engines, schemas, and cluster sizes.

**Impact:** Sizing and schema decisions before go-lives and upgrades are based on measured query performance instead of guesses.

**Architecture:** k6 scenarios → Helm on Kubernetes → Rill + OLAP → Query benchmarks
**Skills:** Performance Testing, k6, Kubernetes, ClickHouse, Benchmarking
**Stack:** k6, Docker, Kubernetes, Helm, Rill, ClickHouse, Apache Druid

## Writing

Index: https://rohithreddykota.com/blog/ · RSS: https://rohithreddykota.com/rss.xml

### Moving HyperLogLog sketches from BigQuery to ClickHouse without the raw data

https://rohithreddykota.com/blog/bigquery-hll-sketches-to-clickhouse-uniqcombined64/ · Published 2026-09-28 · ClickHouse, BigQuery, HyperLogLog, Data Migration, Rust

Summary (from the post): How I converted BigQuery’s ZetaSketch HLL++ sketches into ClickHouse uniqCombined64 states by transplanting the registers directly: no rehashing, no raw rows, and counts within HLL’s own noise.

Key takeaways:
- A HyperLogLog sketch is just an array of per-bucket maximum ranks. BigQuery (ZetaSketch) and ClickHouse (uniqCombined64) store the same array in different binary layouts, so converting between them is a transcode, not a recomputation.
- The two engines hash differently, and that doesn’t matter. A bucket index is an opaque label, and a consistent relabelling leaves everything the estimator reads unchanged.
- The converter runs inside ClickHouse as an executable UDF (a static Rust binary), so a migration is one INSERT … SELECT per partition.
- The encoder reproduces ClickHouse’s own serializer byte for byte. End-to-end counts land within 0.4% of BigQuery’s estimates, which is the difference between the two engines’ bias-correction tables.
- The one rule to respect afterwards: never merge converted sketches with sketches ClickHouse built itself from the same data.

## Skills and stack

Core tools he runs in production every day are listed first in each group.

- **Databases & OLAP:** ClickHouse, Apache Druid, DuckDB (core), plus MotherDuck, Snowflake, Postgres, Redshift, Databricks, Iceberg, Delta Lake
- **Streaming & batch:** Kafka, Apache Beam, Dataflow (core), plus Flink, Spark, Airflow, dbt
- **Infrastructure:** Kubernetes (GKE), Helm, Terraform (core), plus Go operators, Docker, GitOps, GitHub Actions
- **Identity & security:** SSO, SAML, OIDC (core), plus OAuth2, IAM, Row-level security, Bucket policies
- **Languages & APIs:** Python, Go, SQL (core), plus Scala, FastAPI, REST, gRPC, Webhooks
- **Cloud:** Google Cloud, AWS (core), plus GCS, S3, EMR, IAM
- **AI tooling:** MCP, Claude Code, Cursor (core), plus Codex, OpenAI API, Structured outputs, LLM agents

## Education

- M.S., Data Architecture & Management, Northeastern University, Boston, MA (2023 – 2025)
- Executive MBA, Digital Marketing & Analytics, Indian School of Business, India
- B.Tech, Electronics & Communication Engineering, Amity University, Noida, India (2012 – 2016)

## Certifications

- Professional Cloud Architect (Google Cloud)
- Build Infrastructure with Terraform on Google Cloud (Google Cloud)
- Optimize Costs for Google Kubernetes Engine (Google Cloud)
- Develop Google Cloud Network (Google Cloud)
- Implement Load Balancing on Compute Engine (Google Cloud)
- Certification for Apache Airflow (Astronomer)

## Customers

He has taken 50+ enterprise customers live on Rill and run 150+ proofs of concept, mostly in AdTech and marketing analytics. Public Rill customers, as listed on https://www.rilldata.com, include The Trade Desk, Comcast, AT&T, AppLovin, FreeWheel, InMobi, Moloco, Liftoff, MNTN, Chartboost, Verve, Cadent, tvScientific, Invidi, MobileFuse, Kevel, Cognitiv, CloudX, Sabio, Emodo, CreatorIQ, Angel Studios, Deepgram, NOCD, Disco, BlueCargo, Growthcode, Audiohook, Mile, RideAlso, and Wakefit.

## Open source

- [rilldata/rill](https://github.com/rilldata/rill): Open-source, real-time operational BI on ClickHouse, Druid, and DuckDB. Role: Early engineer & contributor.
- [rill-agent-skills](https://github.com/rohithreddykota/rill-agent-skills): Agent skills that let Claude Code and Cursor develop, query, and debug Rill projects. Role: Author. Install: `npx skills add rohithreddykota/rill-agent-skills`
- [rill-clickhouse-incremental](https://github.com/rohithreddykota/rill-clickhouse-incremental): Idempotent incremental ingestion into ClickHouse with Rill, as a runnable example.
- [rill-open-rtb](https://github.com/rohithreddykota/rill-open-rtb): Programmatic-advertising bid logs analyzed with the OpenRTB framework in Rill.
- [the-k8s-operator](https://github.com/rohithreddykota/the-k8s-operator): A CronJob Kubernetes operator in Go, built with Kubebuilder, with a step-by-step guide.
- [k8s-service-discovery-in-prometheus](https://github.com/rohithreddykota/k8s-service-discovery-in-prometheus): Automatic discovery of Kubernetes services as Prometheus scrape targets.

## Frequently asked questions

Full answers: https://rohithreddykota.com/about/#faq and https://rohithreddykota.com/llms-full.txt

- [Who is Rohith Reddy Kota?](https://rohithreddykota.com/about/#who-is-rohith-reddy-kota)
- [What does Rohith Reddy Kota do at Rill Data?](https://rohithreddykota.com/about/#what-does-rohith-reddy-kota-do-at-rill-data)
- [Is Rohith Reddy Kota a forward deployed engineer?](https://rohithreddykota.com/about/#is-rohith-reddy-kota-a-forward-deployed-engineer)
- [What industries does Rohith Reddy Kota specialize in?](https://rohithreddykota.com/about/#what-industries-does-rohith-reddy-kota-specialize-in)
- [How many years of experience does Rohith Reddy Kota have?](https://rohithreddykota.com/about/#how-many-years-of-experience-does-rohith-reddy-kota-have)
- [How does Rohith Reddy Kota deploy an enterprise customer?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-deploy-an-enterprise-customer)
- [How many customers has Rohith Reddy Kota worked with?](https://rohithreddykota.com/about/#how-many-customers-has-rohith-reddy-kota-worked-with)
- [Does Rohith Reddy Kota have product management experience?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-have-product-management-experience)
- [What customer success work does Rohith Reddy Kota do?](https://rohithreddykota.com/about/#what-customer-success-work-does-rohith-reddy-kota-do)
- [How does Rohith Reddy Kota onboard and train customers?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-onboard-and-train-customers)
- [How does Rohith Reddy Kota support customers with technical issues?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-support-customers-with-technical-issues)
- [Does Rohith Reddy Kota set up SSO and security for customers?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-set-up-sso-and-security-for-customers)
- [Does Rohith Reddy Kota work with executives and non-technical stakeholders?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-work-with-executives-and-non-technical-stakeholders)
- [What data migrations has Rohith Reddy Kota done?](https://rohithreddykota.com/about/#what-data-migrations-has-rohith-reddy-kota-done)
- [Does Rohith Reddy Kota do reverse ETL?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-do-reverse-etl)
- [How does Rohith Reddy Kota migrate customers from legacy systems?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-migrate-customers-from-legacy-systems)
- [Does Rohith Reddy Kota handle database and system upgrades?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-handle-database-and-system-upgrades)
- [How much data does Rohith Reddy Kota manage?](https://rohithreddykota.com/about/#how-much-data-does-rohith-reddy-kota-manage)
- [Does Rohith Reddy Kota deploy distributed databases on Kubernetes?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-deploy-distributed-databases-on-kubernetes)
- [How does Rohith Reddy Kota approach database performance?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-approach-database-performance)
- [How does Rohith Reddy Kota keep data infrastructure costs under control?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-keep-data-infrastructure-costs-under-control)
- [Does Rohith Reddy Kota build batch or real-time data pipelines?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-build-batch-or-real-time-data-pipelines)
- [How does Rohith Reddy Kota handle ingestion failures?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-handle-ingestion-failures)
- [How does Rohith Reddy Kota debug production issues?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-debug-production-issues)
- [What AI and LLM work has Rohith Reddy Kota shipped?](https://rohithreddykota.com/about/#what-ai-and-llm-work-has-rohith-reddy-kota-shipped)
- [What internal tooling has Rohith Reddy Kota built?](https://rohithreddykota.com/about/#what-internal-tooling-has-rohith-reddy-kota-built)
- [What projects has Rohith Reddy Kota built?](https://rohithreddykota.com/about/#what-projects-has-rohith-reddy-kota-built)
- [What programming languages does Rohith Reddy Kota use?](https://rohithreddykota.com/about/#what-programming-languages-does-rohith-reddy-kota-use)
- [Which databases does Rohith Reddy Kota work with?](https://rohithreddykota.com/about/#which-databases-does-rohith-reddy-kota-work-with)
- [Which cloud platforms does Rohith Reddy Kota use?](https://rohithreddykota.com/about/#which-cloud-platforms-does-rohith-reddy-kota-use)
- [Does Rohith Reddy Kota drive process change?](https://rohithreddykota.com/about/#does-rohith-reddy-kota-drive-process-change)
- [How does Rohith Reddy Kota work in ambiguous situations?](https://rohithreddykota.com/about/#how-does-rohith-reddy-kota-work-in-ambiguous-situations)
- [What side projects has Rohith Reddy Kota built?](https://rohithreddykota.com/about/#what-side-projects-has-rohith-reddy-kota-built)
- [Has Rohith Reddy Kota worked at an early-stage startup?](https://rohithreddykota.com/about/#has-rohith-reddy-kota-worked-at-an-early-stage-startup)
- [Where did Rohith Reddy Kota study?](https://rohithreddykota.com/about/#where-did-rohith-reddy-kota-study)
- [What certifications does Rohith Reddy Kota hold?](https://rohithreddykota.com/about/#what-certifications-does-rohith-reddy-kota-hold)

## Contact

The best way to reach him is a LinkedIn message: https://www.linkedin.com/in/rohithreddykota/

- LinkedIn (in/rohithreddykota): https://www.linkedin.com/in/rohithreddykota/
- GitHub (@rohithreddykota): https://github.com/rohithreddykota
- Bluesky (@rohithreddykota.com): https://bsky.app/profile/rohithreddykota.com
- X (Twitter) (@rohithreddykota): https://x.com/rohithreddykota
- Blog (whiletrue.live): https://sde.whiletrue.live/