Yash Jangid

I began my journey as a game developer quite some time ago, focusing on creating fantastic games, physics systems, and simulations for three consecutive years. This initial experience instilled in me a profound appreciation for performance, memory management, networking, and the behavior of systems under stress. After a brief hiatus from programming to pursue my academics, I transitioned into backend and infrastructure, delving into fundamental computer concepts. However, I continued to work on low-level projects, developing game engines and simulations. It was during this time that I discovered deep Q learning, which sparked my interest in reinforcement learning. This curiosity eventually expanded to encompass agentic AI systems, which are currently trending, leading me to create systems that integrate my skills.

Learning by Shipping

I gain knowledge by constructing and launching real systems. From my first year, I learned and developed numerous projects. In the following months, I successfully launched two platforms with vastly different results. One was a peer-to-peer file-sharing platform that attracted over 20,000 organic users and managed more than a million requests, teaching me how to build packets, manage buffers, and design for traffic surges and bandwidth limitations. Unlike other platforms, it imposed no restrictions, continuously streaming data to the hard disk at speeds exceeding 200 Mbps, while ensuring operational reliability. The other was a social platform that faced challenges with user retention. This contrast highlighted the importance of robust engineering, but also emphasized that product clarity and understanding user behavior are equally crucial.

What I’m Learning Now

At present, I am concentrating on enhancing my knowledge of AI systems and agentic architectures, particularly focusing on the interactions between models, tools, data, and users in a production environment. In addition to this, I am improving my skills in backend development, DevOps, and system architecture — creating scalable services, deployment pipelines, and infrastructure that can consistently handle real-world workloads.

Let’s Build

I’m always eager to collaborate, particularly in the areas of:

• Practical AI systems and agentic architectures
• Production-ready AI deployment and inference
• Backend-heavy and real-time systems
• Open-source tools close to core infrastructure

Don’t hesitate to get in touch if you’d like to share ideas, work together on a project, or discuss the future direction of real-world AI systems.