Backend SDE II

Evaratus turns enterprise data into datasets, RL environments, and evaluations used in frontier AI research.

Its teams ingest and prepare large data volumes, build benchmarks, and create systems that simulate complex work.

The company is early stage, has about 50 full-time employees, and works with frontier AI labs including OpenAI and Anthropic.

As an engineer at Evaratus, you will own backend systems that ingest, process, and prepare petabyte-scale enterprise data.

You will build distributed data platforms, RL environments, eval harnesses, or research infrastructure. The role requires high agency, sound systems judgment, and fast learning.

You will work on-site with engineering and research teams in Bengaluru.

• Distributed data systems. Build backend services that process petabyte-scale enterprise data.
• Data ingestion pipelines. Ingest, scrub, structure, and store data for downstream research.
• Enterprise data platforms. Turn repeated project work into reliable systems for partner labs.
• RL environments and harnesses. Build environments, eval tooling, or agentic systems based on team fit.
• Zero-to-one architecture. Define processes and technical foundations where no established pattern exists.
• Production quality. Ship quickly while protecting correctness, privacy, and difficult edge cases.

• Built distributed backend systems. You made architecture decisions and owned production outcomes.
• Handled petabyte-scale data. You solved ingestion, processing, storage, or reliability problems at this scale.
• Built beyond standard ETL. You owned backend behavior, failure modes, and platform design.
• Created systems from zero. You set the process, made tradeoffs, and shipped without a mature playbook.
• Learned unfamiliar domains quickly. You turned incomplete context into working software.
• Built agents or eval harnesses. You have relevant exposure to agentic systems, though it is optional.
• Carried broad accountability. You operated as a founder, founding engineer, or end-to-end owner.

• Own the outcome end to end. Take a loosely framed problem through design, delivery, and production.
• Learn while building. Use strong systems foundations to enter unfamiliar research areas quickly.
• Move with the model cycle. Keep iterations short as new models and requirements emerge.
• Protect the edge cases. Pair delivery speed with careful handling of privacy, failures, and correctness.
• Choose tools for the problem. Language matters less than architecture and sound engineering judgment.
• Build reusable platforms. Turn repeated project needs into systems that reduce churn for partner labs.
• Work together on-site. Collaborate with engineering and research teams in the Bengaluru office.

Cash compensation of ₹50L to ₹80L a year. You will work on-site in Bengaluru across backend, data, platform, and research engineering. The role gives you ownership of systems used to create data, environments, and evaluations for frontier AI labs.