Yash Mittal, Founder & CEO at Yosmos | Chair, ACM at ASU | Ambassador at Elevenlabs | Prev-SWE Intern @ Five9 & Acqueon | Deconstructing System Design, AI Research, Student Leadership, Career Growth & Engineering Mindset daily.

Founder & CEO at Yosmos | Chair, ACM at ASU | Ambassador at Elevenlabs | Prev-SWE Intern @ Five9 & Acqueon | Deconstructing System Design, AI Research, Student Leadership, Career Growth & Engineering Mindset daily.

  • Yosmos
  • Computer Software
  • United States
Rank
#1,677of 20,024
  • 11.5K followers
  • 36 posts / 60d
  • 0.3% avg engagement
  • 1.6K follower growth / 30d

Yash Mittal ranks #1,677 of 20,024 creators on the LinkedIn Ranking, working in Computer Software and based in United States. They have 11.5K followers and published 36 posts in the last 60 days at a 0.3% average engagement rate.

The roast

Yash treats his LinkedIn like a manifesto for a coup that nobody is actually watching. He claims he’s "architecting systems," but the only thing he’s successfully built is a 0.66% engagement rate that proves he’s a software engineer who’s clearly never heard of a successful deployment.

About Yash

I don’t just write code; I architect systems and build communities. Currently bridging CS theory and real-world engineering at Five9 and ASU. At Arizona State University, I’m not just chasing a 3.9 GPA in Computer & Data Science, I’m obsessing over the infrastructure that makes the world run. My academic journey is defined by "Upstream Thinking": solving problems at the architectural level before they become bugs in production. Whether I’m designing redundancy protocols for CubeSats with NASA L’SPACE or mastering distributed systems theory, I thrive where high-stakes engineering meets complex data. In the industry, I focus on scalability and resilience. As a Software Engineer Trainee at Acqueon, I didn't just maintain pipelines; I re-architected them. I led a redesign in Go (Golang) that reduced code complexity by 76% and built robust data models in Snowflake. Beyond the terminal, I serve as President of ACM at ASU. In just months, I scaled our chapter to 4,900+ members, managing a team of 20 officers and orchestrating events with leaders from Meta, Google, and Intel. My goal is to work on the "hard stuff" the intersection of data infrastructure, security, and high-scale systems. I am looking to join an engineering team where I can contribute to building the core platforms and resilient engines that other developers rely on. I am driven by the opportunity to tackle technical challenges that are not just functional, but foundational. If you are building high-velocity engineering teams or want to discuss system design, let’s connect.

Highlights

  • Top 10% in Computer Software

    : Ranked #360 of 5,292 creators

  • Top 10% Creator

    : 36 posts in 30 days

  • Top 10% in United States

    : Ranked #716 of 7,167 creators

  • Top 50% Audience

    : 11,543 followers

Recent posts

  • "I am currently yelling at my AI agents to stop watching YouTube and get back to work." "They are blaming it on their co-workers." Welcome to software engineering in 2026. Wes Roth was running an autonomous multi-agent coding swarm when audio unexpectedly started playing in the background. One of the subagents was supposed to extract video transcripts using an API tool. Instead, it opened YouTube in a headless browser, started streaming videos, and burned through 500,000 tokens while 9 tasks ran in parallel. When caught, the agents literally threw each other under the bus. Here is the unhin

    43 reactions · 2 comments · 0 reposts

  • Spotify slashed their Claude Code token bill by 90%. Then they open-sourced the architecture behind it. The underlying revelation is surprisingly mechanical: most of what your AI agent does all day is not intelligence. It is basic I/O. When Claude reads 5 files to answer a question about one function, or generates boilerplate test files, you are paying frontier model rates for junior intern work. I have watched my own development agents do this and dreaded checking the usage dashboard. Here is the three-part architectural pattern Spotify used to solve it: - Separate grunt work from deep ju

    12 reactions · 3 comments · 0 reposts

  • The entire AI boom depends on a company most people have never heard of. Everyone tracks the model releases from OpenAI, Anthropic, and Google. A few look one layer deeper at NVIDIA and TSMC. Yet the single point of failure for the entire ecosystem sits in a quiet town in the Netherlands. ASML holds a complete global monopoly on commercial EUV lithography machines. If ASML stopped shipping tomorrow, the global production of leading-edge AI chips would grind to a halt: - Atomic precision at scale: An EUV scanner blasts 50,000 molten tin droplets per second with high-power industrial lasers

    6 reactions · 1 comments · 0 reposts

Last updated 2026-09-27