Senior Site Reliability Engineer
Akshat Sinha
6+ years building large-scale distributed infrastructure and driving reliability at scale. Research published at USENIX Security, IMC, and ICVGIP. Discovered CVEs in high-profile OSS AI projects. Artifact reviewer for OOPSLA and MLSys 2026.
Education
University of Wisconsin–Madison
MS, Computer Science · 2021–2023
CS Departmental Scholarship · GPA 4.0 / 4.0
IIIT Delhi
BTech, Computer Science · 2014–2018
GPA 8.76 / 10.0
Skills
currently using
Work
Senior Software Engineer, SRE · Rubrik
- Serving as Incident Commander for 2+ years — leading cross-functional response for critical production incidents, coordinating engineering, comms, and executive stakeholders.
- Spearheaded Observability as Code (Terraform) migration for 50+ teams, leading 3 SREs over 4 quarters with reusable alert templates.
- Capacity-planned the company-wide Jenkins cluster, cutting build queue delays by 30% and reducing pod scheduling failures.
- Led design and rollout of a StackStorm-based auto-remediation framework, reducing MTTR via least-privilege automation.
- Built an LLM-powered Slack summarizer with role-specific contexts (manager vs. on-call) and a GenAI documentation translator.
Site Reliability Engineer · TikTok USDS
Video Architecture Team
- Automated new node deployment pipeline using GitOps and Python, reducing end-to-end deployment from 3–4 hours to minutes.
Production Engineer Intern · Meta (Facebook)
Payments Team
- Built a distributed transaction settlement service (Python, MySQL, RPC) to aggregate fintech provider logs, deployed on Meta's container orchestration system.
Site Reliability Engineer · Media.Net (Directi)
Ads Serving Team
- Built and deployed a vulnerability detection system (Wazuh + ElasticSearch), fixing 1000+ system-level issues across the fleet.
- Migrated 10+ legacy applications to Kubernetes/Nomad, cutting infrastructure costs by 50%.
- Enhanced monitoring stack with Prometheus and Grafana to support mission-critical SLAs.
Research Projects
New Strongly Consistent Protocol on Kafka
Implemented a strongly consistent protocol for publishing messages to a Kafka cluster. Achieves performance comparable to weaker consistency models while guaranteeing linearizability. Ongoing research project.
Publications
Yuvraj Patel, Chenhao Ye, Akshat Sinha, Abigail Matthews, Andrea C. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau, Michael M. Swift
25 citations · via Semantic Scholar
Tarun Kumar Yadav, Akshat Sinha, Devashish Gosain, Piyush Kumar Sharma, Sambuddho Chakravarty
65 citations · via Semantic Scholar
Akshay Sethi, Akshat Sinha, Ayush Agarwal, Chetan Arora, Anubha Gupta
18 citations · via Semantic Scholar
Technical Articles
Service
Artifact Reviewing
CVE Discoveries
Shell command injection via unescaped newlines in quote() object .op values
Sandbox escape to arbitrary OS command execution when NodeVM is created with nesting: true
IDOR/BOLA in seven monitor API endpoints allowed unauthorized read, write, and delete operations on sessions, messages, build artifacts, and LLM transaction logs.
Arbitrary file write via path traversal in the v2 file upload API enabled remote code execution through a multipart filename bypass.
Unauthenticated IDOR on the image download endpoint allowed cross-tenant exfiltration via flow_id enumeration and missing ownership checks.
Path traversal in audio transcription endpoint leaking server filesystem paths via error messages
CVE Remediation Reviews
Use-after-free in io_uring subsystem — freed heap chunk could enable arbitrary code execution
NULL pointer dereference and memory leak in io_uring queue initialization, enabling local DoS
Contact
Interested in collaborating or have a question? Send me a message and I'll get back to you.