I build infrastructure platforms that scale, and the agents that run them.
Senior Software Engineer at LinkedIn with 7+ years in infrastructure and distributed systems. I've spent the past 4 years building the control planes and tooling that 10,000+ engineers use to provision and operate 30+ infrastructure platforms, and now I'm layering production agentic AI on top to make them faster, more self-serve, and in some cases fully automated.
The control plane and unified CLI behind a company-wide effort to move all infrastructure provisioning onto a single IaC stack. I architected it to handle authoring, validation, publishing, and full lifecycle management across many platforms, and 10,000+ engineers use it daily. I drove it to production-readiness and cut CLI cold-start from 30s to ~200ms.
Built the core resource-manager layer at the heart of the infrastructure control plane. It's the foundation now powering 30+ infra platforms (MySQL, Couchbase, Temporal, TiDB, and more) with self-serve provisioning.
A streaming framework that auto-rightsizes capacity for a large distributed document store, with pluggable recommendation engines, policy-driven enforcement, and near-real-time utilization tracking. Targets roughly $1.6M/yr of reclaimable capacity.
Built three pillars of research infrastructure at Goldman Sachs: a database-agnostic graph database for fraud detection (Neptune, JanusGraph, Gremlin), a graph search service on Elasticsearch (autocomplete, fuzzy, weighted entity lookups), and a document/coverage store on Cassandra that cut response times ~35% over the legacy system.
An agentic layer over the IaC control plane that takes infrastructure from natural-language intent to running, production-ready resources, so users never touch platform-specific rules or hit failing validations. It's pluggable by design: a new platform comes online by authoring a single rules file, which cut agent setup from about 17 days to 1 day.
Re-architected the manifest model into a multi-file, rule-based structure that closed a file-level ACL security gap and bounded agent context, then migrated 154 services across 90+ teams onto it. Zero outages, one engineer, run end to end by AI agents that raised the PRs, repaired CI, pinged owners, and tracked closure.
Proved Crossplane can manage stateful clusters, not just stateless apps: a ClickHouse control plane covering bootstrap, scale, reshard, red/black failover, backup, and restore via custom CRDs and phase-machine controllers. That de-risks moving teams off bespoke operators and onto a shared platform.
A two-tier oncall system: a user-facing assistant that deflects repetitive queries, and an engineer-facing agent that automates the first 20–30 min of every shift, handling log-anomaly detection, metrics, RCA, and reporting. Backtested on real incidents, it matched human oncall judgment and saved 10–42 minutes per investigation.
Own the authoring, CI, and tooling layers of the IaC control plane, plus the agentic-AI layer that now authors, migrates, and operates infrastructure end to end. Technically lead a small engineering pod.
Built the core resource-manager layer of the infrastructure control plane, now powering 30+ infra platforms, plus a streaming quota framework (~$1.6M reclaimable) and an oncall copilot. Migrated the MySQL platform onto a modernized control plane solo.
Built distributed graph, search, and document infrastructure for fraud detection and ML-powered research.
$ whoami
Amandeep Srivastava, Infrastructure & Platform Engineer
$ cat journey.log
4 years at LinkedIn building the control planes behind the systems, now powering 30+ infrastructure platforms.
Before that, graph, search, and document infrastructure for fraud detection at Goldman Sachs.
Now layering agentic AI on top, so 10,000+ engineers provision and operate infra by intent, not toil.
$ echo $EDUCATION
IIT (ISM) Dhanbad · B.Tech CS (Honors) · Class of 2019
$ echo $INTERESTS
distributed-systems, control-planes, infrastructure-as-code, agentic-AI, AI-for-infra
$ _