Backend SDE 3

Evaratus turns enterprise data into research infrastructure for frontier AI labs.

It builds large datasets, RL environments, evaluation harnesses, benchmarks, and systems that simulate complex work.

The early-stage team has about 50 employees and works with frontier AI labs including OpenAI and Anthropic.

As SDE 3 at Evaratus, you will own distributed backend and data systems used in frontier AI research.

You will build petabyte-scale ingestion and processing systems from the ground up. The role requires strong distributed systems craft, broad backend judgment, and the ability to learn new domains quickly.

You will own workstreams, mentor engineers, and unblock the team.

• Distributed backend systems. Build reliable systems that process large, complex datasets.
• Petabyte-scale data flows. Own ingestion, transformation, storage, and movement across the platform.
• Zero-to-one system builds. Set up new services, processes, and technical foundations from scratch.
• Research data infrastructure. Support datasets and environments used by frontier AI labs.
• Technical workstreams. Take systems from an open problem through production operation.
• Engineering mentorship. Guide junior engineers, review their work, and remove delivery blockers.

• Built distributed systems hands-on. You made key design calls and carried a backend system into production.
• Handled petabyte-scale data. You can explain the tradeoffs behind storage, processing, and reliability.
• Built systems from zero to one. You created the first workable process where little infrastructure existed.
• Owned more than ETL pipelines. Your work included backend behavior, reliability, or distributed processing.
• Learned unfamiliar domains quickly. You became effective while the problem and technical context evolved.
• Led work while still building. You mentored engineers, removed blockers, and retained direct system ownership.

• Own the system end to end. Carry an open problem through design, implementation, and production operation.
• Build from first principles. Create practical foundations when established systems and processes are limited.
• Choose tools for the problem. Language matters less than sound distributed systems decisions.
• Learn while building. Pick up agentic AI, harnesses, or evaluations as the work requires.
• Judge engineers by their work. Strong systems built in any company carry more weight than brand names.
• Work together on-site. The engineering team collaborates in person five days a week.
• Raise the team's capacity. Mentor junior engineers and unblock delivery without leaving hands-on work.

Cash compensation of ₹85L to ₹1.2Cr a year. The compensation structure and reporting line are being finalized. You will own distributed systems that support petabyte-scale data and frontier AI research.