Karthik Sunkishala profile photo 1 on Tal

Karthik Sunkishala

Building AI Platforms | IIT Bhubaneswar CSE’21

Bengaluru

5+ years of work experience

aion
Oct 20251 yr 1 moPresent

Tech Lead for the Ai Platforms Charter. Solving interesting problems across inference optimisations, fine-tuning pipelines, productionising enterprise agents and their evals
Inferless
Nov 20232 yrsOct 2025

Bengaluru

Founding Engineer. Built a reliable & fast serverless GPU Inference Platform for LLMs Reduced Cold Starts from minutes to seconds at scale
Dream11
Jul 20235 mosNov 2023

On-site

Built unified Machine Learning Platform @Dream11 1. Held ownership of core components of the platform. 2. Led multiple POCs with Ray, GPUs and GenAI. 3. Designed and Implemented a multi-cloud architecture on AWS & GCP 4. Worked closely with data scientists in scaling model inference to a Million concurrency
Dream11
Jun 20212 yrs 1 moJun 2023

Mumbai, Maharashtra, India

Building a unified ML Platform where I have primarily worked on designing, developing and scaling ML Compute and ML Inference Systems. Worked extensively with Python, Kubernetes, Ray, Golang and AWS
GE Healthcare
Jun 20203 mosAug 2020

What I'm currently working on

I am leading the AI platform services team at Aion, focusing on managed inference and LLM optimization frameworks.

What's hard for others but easy for me…

Architecting distributed compute platforms for ML training is straightforward for me because I enjoy solving for scale and latency.

Karthik’s AI Stack

Braintrust
Braintrust
Bubble AI
Bubble AI
Modal
Modal

Education

Indian Institute of Technology, Bhubaneswar

BTech - Bachelor of Technology, Computer Science

Indian Institute of Technology, Bhubaneswar

20174 yrs2021

Join Karthik and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Karthik Sunkishala on Tal

Headline

Building AI Platforms | IIT Bhubaneswar CSE’21

About

🚀 Building Scalable AI/ML Systems Most recently, I was an SDE 3 at Inferless, where we were building serverless model inferencing. My work involved architecting and developing scalable components/platforms from scratch to support custom models efficiently. Previously, I was an SDE 2 at Dream11, where I contributed to building a unified Machine Learning platform. 🔹 Proficient in: Backend Systems, Kubernetes, Distributed Systems, ML/AIOps, 🔹 Programming Languages: Python, Golang, Rust, C 🔹 Cloud Expertise: AWS | GCP | Azure My passion lies in AI/ML systems, and I am committed to pushing the boundaries of what's possible in this exciting domain.

Location

Bengaluru

AI stack

Braintrust

Bubble AI

Modal

Work experience

Engineering at aion, Oct 2025 - Present · 1 yr 1 mo, Tech Lead for the Ai Platforms Charter. Solving interesting problems across inference optimisations, fine-tuning pipelines, productionising enterprise agents and their evals

SDE 3 at Inferless, Nov 2023 - Oct 2025 · 2 yrs, Bengaluru, Founding Engineer. Built a reliable & fast serverless GPU Inference Platform for LLMs Reduced Cold Starts from minutes to seconds at scale

SDE 2 at Dream11, Jul 2023 - Nov 2023 · 5 mos, On-site, Built unified Machine Learning Platform @Dream11 1. Held ownership of core components of the platform. 2. Led multiple POCs with Ray, GPUs and GenAI. 3. Designed and Implemented a multi-cloud architecture on AWS & GCP 4. Worked closely with data scientists in scaling model inference to a Million concurrency

SDE-1 at Dream11, Jun 2021 - Jun 2023 · 2 yrs 1 mo, Mumbai, Maharashtra, India, Building a unified ML Platform where I have primarily worked on designing, developing and scaling ML Compute and ML Inference Systems. Worked extensively with Python, Kubernetes, Ray, Golang and AWS

EID intern at GE Healthcare, Jun 2020 - Aug 2020 · 3 mos

Education

BTech - Bachelor of Technology, Computer Science, Indian Institute of Technology, Bhubaneswar

Profile highlights

What I'm currently working on: I am leading the AI platform services team at Aion, focusing on managed inference and LLM optimization frameworks.

What's hard for others but easy for me…: Architecting distributed compute platforms for ML training is straightforward for me because I enjoy solving for scale and latency.