I enjoy taking ambiguous, messy problems and turning them into simple, reliable systems. My work usually sits at the intersection of product intent, system design, and real-world constraints. I prefer boring tech at scale, obsess over system boundaries, and optimize for correctness before cleverness.

• Break vague product ideas into clear, solvable system problems
• Make pragmatic tradeoffs under performance, cost, and reliability constraints
• Design solutions that are easy to reason about, operate, and evolve
• Treat performance, infra, and UX as parts of the same system

Currently deepening my understanding of distributed systems, especially around inference pipelines, failure handling, and scalability tradeoffs.

🧩 Distributed Inference Engine
🔗 https://github.com/ishanjain1502/distributed-inference-engine
Built a distributed infrastructure to serve LLM inference for public-facing applications.

• Designed the system end-to-end, from request flow to worker orchestration and session lifecycle
• Chose session termination over recovery on worker decode failure, favoring fast restarts and predictable behavior over complex recovery paths
• Focused on clear system boundaries and failure modes rather than over-engineered abstractions

💸 High-performance Discount Engine (WASM · Shopify Functions)
Worked on a revenue-critical discount system operating in a constrained runtime.

• Identified inefficiencies in how discount data was structured and processed across features
• Re-architected the data model and execution flow before optimizing performance
• Ported critical paths to Rust + WASM to improve execution speed and reliability in checkout flows

⚡ Web Vitals Optimization (LCP / CLS)
🔗 https://ishanjain.hashnode.dev/web-vitals-and-their-optimizations
Solved real performance problems in a production application with direct SEO and UX impact.

• Improved Core Web Vitals, leading to better search ranking and overall user experience
• Established internal standards for performance work across the organization
• Approached frontend performance as a system-level problem, not isolated UI tweaks

Languages: TypeScript, JavaScript, Rust, Python
Systems & Infra: Docker, AWS, GraphQL, WebAssembly
Data: PostgreSQL, MongoDB
Product & UI: React, Next.js

(I care more about why a tool is used than which tool is used.)

• Email: ishanjain1502@gmail.com
• LinkedIn: https://linkedin.com/in/hey-ishan