Shivam Ashtikar

SDE II · AI Infrastructure Specialist

Juspay · 7+ years experience

Software Development Engineer (SDE II) with 7+ years of experience architecting scalable systems, leading cross-functional teams, and overseeing full-cycle development. Expertise spans Payment Systems, Data Engineering, Infrastructure Automation, and advanced LLM Inference Engines.

LLM Inference GPU Infrastructure Functional Programming Team Leadership

How I Work

From idea to production, emphasizing measurement, automation, and continuous optimization across AI infrastructure and full-stack systems.

90%

Bug reduction via Canary

700+

Daily active LLM users

Years of experience

Ship fast · Measure impact · Iterate continuously

Journey

SDE II @ Juspay

Mar 2023 – Present

Current

7+ years across AI Infrastructure, Full-Stack Platforms, and Payments

Xyne - AI Infrastructure

LLM Inference & MLOps · 2025 – Present

→ Spearheaded AI Infrastructure team deploying open-weight LLMs on 8x H200 GPU nodes with vLLM and SGLang
→ Patched vLLM for Kimi K2.5 compatibility, achieving production deployment within one week of release
→ Led cross-functional teams (Frontend, Backend, QA) to deliver the open-source Xyne platform

vLLM SGLang LiteLLM VespaDB Ray H200 GPUs

Euler Dashboard

Full-Stack Platform Leadership · 2023 – 2025

→ Led Backend and Frontend teams, mentoring engineers and driving technical excellence
→ Reduced production bugs by 90% through Canary deployment strategy
→ Integrated enterprise features: Slack API, OAuth 2.0, Passkey, and Google SSO

Haskell ReScript React Jenkins

SDE I

Jun 2021 – Mar 2023

Led Payment Page engineering team. Reduced PureScript compile time by 6× through strategic refactoring. Built DevTools application for payment flow verification.

6× faster compile times

PureScript React Jenkins

Associate SDE

Jul 2019 – Jun 2021

Full-stack payments development. Improved Android render time by 66% with pre-rendering. Migrated data pipeline from Python to Haskell using Streamly.

66% render improvement

Android Python Haskell

FP Developer Intern

Dec 2018 – Feb 2019

Enhanced cross-platform framework performance by 10% through JS-Java bridge optimization.

10% performance gain

JavaScript Java Android

Technical Toolkit

vLLM

High-throughput LLM serving inference engine

SGLang

Structured generation for LLMs

LiteLLM

Unified LLM proxy (700+ users)

LMCache

KV cache optimization

VespaDB

Vector database for RAG systems

Ollama

Local LLM deployment

Education

Bachelor of Engineering in Computer Engineering

Pune University – Government College of Engineering and Research, Avasari (kd)

CGPA: 8.06 | 2015 – 2019

Tools I've selfhosted

Open resources I maintain for the developer community. Free utilities and services hosted on my infrastructure.

FreshRSS IT Tools Browse Workbench

Always curious, always building — let's connect!

GitHub LinkedIn

[email protected]