Headline

Senior Software Engineer | Building Large-Scale Platform Products | Distributed Systems | AI Infrastructure

About

I build products that solve large-scale engineering problems. Over the past 7+ years, I've worked across cloud infrastructure, financial platforms, and enterprise systems, designing distributed products that process millions of transactions, analyze terabytes of telemetry, and support critical operations. At Oracle Cloud Infrastructure, I proposed and led the development of Pulse, a distributed fleet health platform monitoring 3M+ GPU accelerators and bare-metal servers across more than 100 cloud regions. I also led the design of OCI's observability platform, processing over 5 TB of telemetry daily to help engineering teams diagnose production incidents faster. Previously at Acko, I led the architecture and evolution of Finacko, the company's distributed financial platform processing more than 10 million transactions per day, while building internal platform products that improved developer productivity and operational reliability. I'm particularly interested in: • Distributed Systems • Platform Products • Backend Architecture • Event-Driven Systems • Observability • AI Infrastructure I'm currently going deep on GenAI inference infrastructure — model serving, distributed inference, and the systems problems of running frontier-scale models economically. Open to conversations on distributed systems, GPU infrastructure, and AI infrastructure at scale.

Location

Bengaluru

AI stack

claude code

codex

hugging face

langchain

mcp

Browser Use

Work experience

Senior Member of Technical Staff at Oracle, Jun 2024 - Present · 2 yrs 3 mos, Bengaluru, • Designed and built Pulse, a distributed hardware health monitoring service tracking 300,000+ bare-metal servers and GPU accelerators per region across 100+ global regions, covering H200, GB200, GB300, and MI355X — OCI's first line of detection for hardware failure at fleet scale. • Built a purpose-designed observability platform processing 5TB/day of operational data, giving engineering teams real-time fleet visibility during large-scale compute events and critical incidents.

Software Development Engineer - 2 at Acko, Oct 2022 - May 2024 · 1 yr 8 mos, Bengaluru, Accounting System: Designed and built a scalable automated ledger that maintains all the financial transactions performed during the life-cycle of a policy and can handle 10M+ transactions in a day. Completely removed manual dependency by building automatic report generation for the finance and underwriting team while minimizing cost by 80%. Config Module: Designed and developed a module on top of BigQuery that overcomes the query character limit of 1024K characters and runs in the BigQuery infrastructure without introducing additional costs and increasing noticeable query response time. Dead-Letter Queue Retry System: Designed and developed a scalable DLQ retry architecture that runs on Apache-Airflow. DLQs can be added/removed based on configs. We managed to overcome the 14 days retention period and run infinite retries on scale.

Software Development Engineer - 1 at Acko, Aug 2021 - Sep 2022 · 1 yr 2 mos, Bengaluru, Karnataka, India, Worked on an automated ledger that can easily handle 1M+ financial transactions per day. This automated ledger tracks every financial transaction made during the lifecycle of a policy. Data is pushed into GBQ using Kafka. Built directed acyclic graphs to handle complex workflows using Apache Airflow. So instead of running manual scripts for policy cancellation, migration, etc Airflow DAG can be scheduled and executed.

Software Engineer at Amdocs, Aug 2019 - Jul 2021 · 2 yrs, Pune, Maharashtra, India, Worked as a full-stack engineer on the core product Ordering Management System that manages the life-cycle of an e-commerce order. I was involved in design, development, testing, deployment, and production support. Built and scaled reliable critical services related to payment and credit checks using service-oriented architecture. Improved complexity of the data migration tool that is used to validate migrated data before feeding into OMS by 40%. Reduced API latency for Multi-Channel Self Service by 20% by introducing various compression techniques, and caching that resulted in a much better user experience for the chatbot.

Education

Bachelor of Engineering - BE, Computer Science and Engineering, Institute of Engineering & Technology DAVV, Indore

Higher Secondary, Mathematics, Saraswati Higher Secondary School, Pandhurna

Public links

github: https://github.com/spd123

medium: https://medium.com/@shivamdongre

leetcode: https://leetcode.com/spd123

codeforces: https://codeforces.com/profile/spd123