I'm a Staff Research Engineer at Tejas Networks, where I spend most of my days doing intense hands-on work somewhere in the space between research, engineering, and standardization. A good part of that time goes into 3GPP RAN1/RAN2 standardization, where the conversations tend to revolve around integrated sensing and communications (ISAC), sensing assisted communication (SAC), positioning, AI-RAN, RAN digital twin. The less glamorous part, and probably the most useful, is turning ideas into intellectual property that can survive both simulation and standardization and hopefully can be reduced to practice someday.

What time is left, including weekends, tends to go toward work that sits closer to the hardware. I help build and maintain a CUDA-C++/Python platform for large-scale wireless system simulations. That means spending a lot of time thinking about memory movement, arithmetic intensity, GPU throughput, and where performance disappears when it shouldn't. Much of the effort goes into writing optimized CUDA kernels and real-time GPU algorithms for AI-RAN. The goal is simple: make simulations faster, more scalable, and faithful to the systems being modeled. Getting there, as usual, is less simple.

I grew up in a village near Ambala, in North Haryana, in a farming family. I loved the farms, the cattle, and mathematics in roughly equal measure. Most afternoons until sunset went to grazing buffalos and working the fields alongside my family. It was a simple life, and a grounding one. The real turning point was IIT Kanpur, where I completed a master's in Signal Processing and Communications, met some genuinely kind and capable people, and was advised by Prof. Aditya K. Jagannatham. That time and supervision shaped most of what I went on to do.

After IITK I joined Intel's SDG group in Bangalore, but left fairly quickly. The work wasn't what I had hoped for. In October 2018 I joined CEWiT (Centre of Excellence in Wireless Technology), where I found on cutting edge research, development and standardization work related to 5G networks. I contributed to 3GPP standardization on positioning from 2018 through late 2021. I focused on the physical layer and MAC-Scheduler side of 5G, mostly positioning, sidelink, and AI-ML. That meant designing every single positioning algorithm for the 5G physical layer, building and maintaining a system simulator covering 3GPP releases 15 through 18, and developing AI-ML models for channel estimation, symbol demodulation, LDPC decoding, and inter-frequency measurement prediction.

In December 2021 I left CEWiT to start Gigayasa Wireless. The idea was to build a GPU-scalable, 3GPP-compliant simulation tool for 5G/5G-Advanced/6G networks with native AI-in-RAN support, capable of running end-to-end system simulations within an hour. I began by implementing the 38.xxx series standards in Python, line by line, using that to validate the tool against real-world 5G scenarios. By early 2022 I was leading a team, turning that work into production grade software for the ease of usage, deployment and support and integrating AI-ML models for beam prediction, CSI compression and reconstruction, and positioning. We couldn't scale the business to where we had hoped, and eventually wound down.

In August 2024 I joined Tejas Networks, picking up where I had left off: system simulations and 3GPP standardization for 5G-Advanced and 6G.

I live in Bengaluru. Outside of work I cook, read, and write occasionally on the blogs page about telecom, AI, quantum communications, GPU scaling, C/C++ CUDA Kernels, system simulation, and the occasional lighthearted take on startups.

The fastest way to reach me is LinkedIn. I'm also on GitHub and Google Scholar. If email's your preference, send a note to vikram.singh.anttal@gmail.com. My brief technical description can be found in my CV (PDF)