Harvey
About this role
This is the infrastructure layer under Harvey's legal AI platform, and the posting gives real numbers: billions of prompt tokens and millions of daily requests. The work splits between building new infrastructure systems and hardening what already exists as the company adds products, regions, customers and usage. Three specifics stand out. You architect multi region deployment strategies that satisfy strict data residency rules for global enterprise customers, which is a real constraint in legal software. You build observability with granular SLA monitoring, burn rate alerts and token attribution for cost tracking, so LLM spend is measurable per customer. And you lead the evolution of CI and CD pipelines to keep developer velocity up. Apply if you have 6 or more years in infrastructure or platform engineering and have scaled large distributed systems.
Who this is for
Requirements as published:
- 6+ years in Infrastructure Engineering or Platform Engineering in a production environment.
- A long track record building and scaling complex, large scale distributed systems.
- Deep proficiency with cloud infrastructure platforms. Azure is preferred, and the posting says GCP or AWS experience transfers well.
- Strong fluency in infrastructure as code: Terraform, Pulumi or CloudFormation.
- Solid understanding of Kubernetes, container orchestration, networking and cloud security at scale.
- Experience with observability tools such as Datadog and Sentry, and incident response practice with PagerDuty or Incident.io.
- Strong programming skills in Python, Go or similar.
The actual day to day:
- Architecting multi region deployment strategies that meet strict data residency requirements for global enterprise customers.
- Building observability infrastructure with granular SLA monitoring, burn rate alerts and detailed token attribution for cost tracking.
- Leading the evolution of CI and CD pipelines to improve developer velocity while holding production stable.
- Balancing new system building against operational excellence on infrastructure processing billions of prompt tokens and millions of daily requests.
Location and office reality:
- Bengaluru. As with Harvey's other India postings, the board flags remote while the role sits with the Bengaluru team. Published at the real location.
Honest fit guidance:
- Data residency across regions is a genuinely hard problem and it is named first in the responsibilities. If you have done it, lead with it.
- Token attribution for cost tracking is an AI era infrastructure skill that barely existed three years ago. Nobody expects a long history in it, but show you understand why it matters.
- Azure preferred again, same as the SRE role. Be explicit if you are coming from AWS.
- 6+ years in Infrastructure Engineering or Platform Engineering in a production environment.
- A long track record building and scaling complex, large scale distributed systems.
- Deep proficiency with cloud infrastructure platforms. Azure is preferred, and the posting says GCP or AWS experience transfers well.
- Strong fluency in infrastructure as code: Terraform, Pulumi or CloudFormation.
- Solid understanding of Kubernetes, container orchestration, networking and cloud security at scale.
- Experience with observability tools such as Datadog and Sentry, and incident response practice with PagerDuty or Incident.io.
- Strong programming skills in Python, Go or similar.
The actual day to day:
- Architecting multi region deployment strategies that meet strict data residency requirements for global enterprise customers.
- Building observability infrastructure with granular SLA monitoring, burn rate alerts and detailed token attribution for cost tracking.
- Leading the evolution of CI and CD pipelines to improve developer velocity while holding production stable.
- Balancing new system building against operational excellence on infrastructure processing billions of prompt tokens and millions of daily requests.
Location and office reality:
- Bengaluru. As with Harvey's other India postings, the board flags remote while the role sits with the Bengaluru team. Published at the real location.
Honest fit guidance:
- Data residency across regions is a genuinely hard problem and it is named first in the responsibilities. If you have done it, lead with it.
- Token attribution for cost tracking is an AI era infrastructure skill that barely existed three years ago. Nobody expects a long history in it, but show you understand why it matters.
- Azure preferred again, same as the SRE role. Be explicit if you are coming from AWS.
Apply on company site
Opens jobs.ashbyhq.com, the employer's own application page. Applying is always free.