Headline

SRE @ Netradyne · building AI agents so PagerDuty doesn’t wake me up at 3AM · AI × Infra × Cloud · occasional over-thinker

About

I build systems that investigate incidents before anyone has to. Technically I'm an SRE. Realistically, I spend most of my time convincing AI agents to do the detective work that used to require a very tired engineer and an unhealthy relationship with grep. Here's what that's looked like: → TraceLens — an AI agent that autonomously digs through logs, metrics, and alerts and tells you what broke and why. Not a dashboard. Not a chatbot. An agent that actually investigates. Shipped it org-wide. It works. → AlertFlow — real-time alert correlation across 150+ services. Because 47 simultaneous alerts about the same thing isn't 47 problems — it's one problem and a terrible Tuesday. → A natural-language log analysis tool so engineers can ask "what was broken at 2:47PM?" instead of writing grep one-liners they'll regret later. Stack I live in: Python · FastAPI · React · AWS · Azure AI Foundry · Docker · K8s · Prometheus · Grafana · Elasticsearch I like working with people who take the work seriously but not themselves. If you're building something at the edge of AI × infrastructure — or just want to talk shop — say hi. I read every message. I respond to most of them. The bar is just: don't open with "hope this finds you well."

Location

Bengaluru

AI stack

antigravity

claude code

codex

gemini

mcp

openai

Work experience

Software Engineer - SRE at Netradyne, Feb 2026 - Present · 7 mos, The job got more interesting. Now I architect the AI systems instead of just shipping them. - Leading reliability and observability for large-scale distributed systems powering real-time AI products. Fancy way of saying: I make sure things don't fall over, and when they do, we know exactly why in minutes not hours. - Architecting internal platforms that automate incident detection, alert correlation, and root-cause discovery. The goal is a world where the system explains itself and nobody has to play detective at midnight. - Resident AI-tooling lead for the SRE org. If it involves an LLM doing something actually useful in production, it probably came through me.

Associate Site Reliability Engineer at Netradyne, Jul 2025 - Feb 2026 · 8 mos, This is where things got fun. - Built TraceLens from scratch — an agentic AI system that autonomously queries logs, metrics, and alerts, reasons over the data, and surfaces root causes. Full stack, solo, shipped org-wide. FastAPI backend, React + SSE streaming frontend, native LLM tool-calling loop. No agent frameworks. Just the spec and stubbornness. - Designed incident detection and escalation pipelines integrated with on-call tooling — so the right person gets paged with context, not just a wall of red. - Containerized and deployed internal tooling at org scale with Docker and Kubernetes on AWS. It scales. I'm unreasonably proud of it.

Cloud Intern at Netradyne, Feb 2025 - Jul 2025 · 6 mos, Bengaluru, Started as an intern. Left having built a platform covering 150+ services. Not bad for six months. - Built AlertFlow — real-time alert correlation across distributed services. Turned alert storms into structured triage queues. Made on-call meaningfully less miserable. - Built and maintained monitoring coverage across Grafana, Datadog, Prometheus, and CloudWatch. At one point I had all four open simultaneously and felt nothing. - Automated incident detection and escalation workflows — reduced MTTR through log-driven diagnostics and structured RCA. The MTTR went down. The sleep quality went up.

Research Associate Intern at Nebulixus, Jun 2024 - Jul 2024 · 2 mos, India, Applied machine learning for predictive economic modeling using regression & clustering techniques. Conducted NLP-based sentiment analysis on economic policies. Developed interactive dashboards with Power BI to visualize economic trends & inflation patterns. Presented findings to stakeholders, contributing to data-driven decision-making.

Education

Bachelor of Engineering - BE, Artificial Intelligence and Machine Learning, New Horizon College of Engineering

Public links

leetcode: https://leetcode.com/KrishAryan

github: https://github.com/KrishAryan12