Shivam Chawla

Story

My first patent was about heat. SSDs throttle when they hit 85–95 °C and Western Digital wanted a smarter way to keep delivering throughput when the silicon was hot. I co-invented the framework that did it. +50 MBps. USPTO US11508416B2.

Now I do the same thing for LLMs. At Rising Team, I'm a senior engineer on the team behind aRTi (the AI Leadership Trainer) — turning LLMs into products that help managers lead better teams.

Used by leaders at Adobe, Google Cloud, Meta, Microsoft, LinkedIn, Disney, Cisco, Visa, Indeed, Bank of Hawaii, Yahoo, DXC, Hershey's, Aramark. Up to 33% retention lifts, 20–100% manager-effectiveness gains, 60–200% eNPS improvements. At Bank of Hawaii alone: 400+ sessions across 1,950 employees in 40 branches; 88% retention lift over 18 months.

The systems work shows up everywhere I touch LLMs — latency budgets, predictive throttling, recovery paths. AI is the same problem in a different costume. Before this I led an infra team of 7 across 5 OEM partners (Dell, HP, Asus, Acer, Lenovo), interned at Meta automating Facebook data-center decommissions, and did my MS at UC San Diego (GPA 4.0/4.0) where I researched LLM-based movie recommender systems with Prof. Julian McAuley using SASRec + ChatGPT.

Patents, education & publications →

Timeline

  1. 2023 — present

    Senior Software Engineer Rising Team

    Remote · CA

    AI platform team — aRTi AI Roleplay, AI Trainer, Mini Kit Framework, Realtime Framework.

  2. 2022

    Front End Intern Meta (Facebook)

    Menlo Park, CA

    React + PHP automation framework for Facebook data-center decommission planning. Delivered 3 weeks ahead of deadline.

  3. 2021 — 2023

    MS Computer Science UC San Diego

    San Diego, CA

    GPA 4.0/4.0. TA across 2 academic years. Researched LLM-based recommender systems with Prof. Julian McAuley (SASRec + ChatGPT).

  4. 2019 — 2021

    Engineer · Infra Tech Lead Western Digital

    Bengaluru, India

    SanDisk Extreme Portable SSD line. Thermal/resource-manager framework (+50 MBps at 85–95 °C). Led infra team of 7 across 5 OEM partners (Dell, HP, Asus, Acer, Lenovo).

  5. 2018 — 2019

    Engineering Intern (×2) Western Digital

    Bengaluru, India

    Ungraceful Shutdown framework + automated Python latency-test harness. Coverage uplift 50% → 90% on Infra/FS modules.

  6. 2015 — 2019

    BE Computer Engineering Thapar Institute

    Patiala, India

    GPA 9.27/10. Co-authored IEEE ICMLDS 2018 paper on three-phase transformer fault detection (85.65% ensemble accuracy).

Stack

LLMs·LangChain·Generative AI·RAG·Prompt Engineering·OpenAI·Anthropic·Gemini·Pinecone·ElevenLabs·Python·Django·React·TypeScript·Next.js·PostgreSQL·Node.js·Realtime Systems·System Design·Recommender Systems·Machine Learning·C++·Embedded Systems·Firmware·NVMe·AWS·Docker·Kubernetes·Git·Mentoring·Public Speaking

Outside work

Chess board

Ardent chess fan.

Same systems I think about at work — opening prep is prompt engineering, the middlegame is latency, the endgame is recovery. Same problem, different costume.

Lichess · ~1700 blitz · slowly climbing