Sai ManojLinkedInGitHub

I managed the systems that deliver 100M webhook events a day.

  1. Event
  2. Queue
  3. Delivery service ×3

    30% fewer failed deliveries

  4. Customer endpoint

    25% lower latency

Experience

Webhooks Platform

LinkedIn, Senior Software Engineer

Owner of three webhook delivery services that carry about 100 million events a day.

Sole owner and the DRI engineer for three webhook delivery services that carry about 100 million events a day. I also maintain the authentication middleware on the API Gateway, which handles about 1 billion requests a day.

Designed a new retry mechanism that cut failed deliveries by 30%, and moved webhook processing to a new framework that cut latency by 25%.

Built a monitoring dashboard and debugging tools for on-call, and shared GenAI skills and plugins that speed up root-cause analysis across teams.

Defender

Microsoft, Software Engineer II

Cut a mobile security app's memory use by 30%.

Led the integration of a partner VPN into Microsoft Defender, a mobile security app, and cut its memory use by 30% on end-user mobile devices.

Built telemetry and crash reporting for production monitoring, and an on-prem lab to validate anti-phishing detection. Wrote cross-platform test automation in Python and bash for weekly regression runs, and mentored junior engineers.

Memory use, before and after

Drawn to scale: 30% lower.

Application Gateway

Microsoft, Software Engineer II

Led Header Rewrite on a web traffic load balancer.

Application Gateway is a web traffic load balancer. I led the Header Rewrite feature, which lets customers change header values as traffic is routed. I defined the public API contract, took it through design approval, and built the backend logic to run rewrites at scale on a busy gateway.

Took on-call rotations, wrote troubleshooting guides, and mentored four new teammates.

A header rewrite, illustrated

  1. Backend sendsLocation: http://backend-1.internal/cart
  2. RuleReplace the backend host with the public host
  3. Client receivesLocation: https://shop.example.com/cart

Network Monitoring

Microsoft, Software Engineer

Wrote the agent that checks network paths and connectivity.

Wrote the agent for Service Connectivity Monitor in C and C++. It sends pings and traceroutes to check network connectivity and path health. I drove the agent's design and set up its unit and end-to-end tests.

Built the backend for ExpressRoute monitoring with microservices and MapReduce, to process large volumes of network path data, and did the relational data modeling.

A traceroute, illustrated

  1. Local router2 ms
  2. Network edge9 ms
  3. Transit18 ms
  4. Destination31 ms

Rate limiter

A crucial part of Gateways that I built

The demo is a token bucket, the idea behind rate limiting at a gateway: five tokens, and a new one every two seconds.

Clients
API GatewayAuthentication

Is quota available for this user?

Rate limiter

5 of 5 tokens left

Behind the gatewayBackend services

429 Too Many Requests: over the limit, answered at the gateway

Every request reaches the API Gateway, which checks authentication and then asks the rate limiter: is quota available for this user? If yes, the request goes on to the backend. If no, the gateway answers 429 Too Many Requests and the backend never sees it.

    This demo needs JavaScript. It sends tickets through the rate limiter, which admits five and then answers 429 Too Many Requests until a new token arrives.

    Tech stack

    Backend

    • Java, Python, C#The languages I work in
    • C and C++A network monitoring agent at Microsoft
    • KafkaThe API Gateway and the webhooks platform at LinkedIn
    • Microservices, MapReduceExpressRoute monitoring at Microsoft

    Cloud

    • Azure, AWS, CloudflareWhere I deploy and run services
    • GitHub ActionsCI/CD pipelines I write and fix
    • Load balancingApplication Gateway at Microsoft

    Data

    • PostgreSQLRelational databases and data modeling
    • RedisIn-memory data store
    • DynamoDBNoSQL database

    AI agent tooling

    • GenAI skills and pluginsRoot-cause analysis and alert correlation for webhook workflows

    I'm a senior software engineer with about nine years of backend experience, most of it on services that run on public cloud and handle a lot of traffic.

    I've worked on a web traffic load balancer, network monitoring and a webhooks platform. Along the way I've mentored new hires and interviewed candidates.

    These days I'm building GenAI skills and plugins that help engineers debug faster.

    If you're hiring, or building something and want another engineer to talk to, say hello on LinkedIn.

    Sai ManojSenior software engineer