✨ Kindly view on desktop for the full interactive architectural experience & simulators.
Sai Jyothi Kumar
Sai Jyothi Kumar
Full-Stack & AI Systems Engineer
Aesthetic Developer Workspace Setup
Scalable Distributed Systems & AI Systems

Architecting High-Scale Platforms & Agentic AI Systems.

Full-Stack Software Engineer with 3+ years of experience scaling enterprise platforms to 400,000+ associates at Cognizant, reducing token costs by 90%, and building sub-15ms applications using TypeScript, Next.js 15, FastAPI, and WASM.

400,000+ Enterprise Associates Scaled
90% Token Cost Overhead Cut
< 15ms Indexed Query Latency
3+ Yrs Full-Stack & AI Systems
TypeScript 90% Token Reduction Architecture Python FastAPI Sub-15ms Indexed Query Latency React 18 Next.js 15 Socratic AI Multi-LLM Orchestration SQLite3 (WASM) Docker & CI/CD 400,000+ Enterprise Scale @ Cognizant ASP.NET Core Angular PostgreSQL Redis Zero-Defect Production Releases AWS GenAI Practitioner 2026 Server-Side Prompt Injection Defense TypeScript 90% Token Reduction Architecture Python FastAPI Sub-15ms Indexed Query Latency React 18 Next.js 15 Socratic AI Multi-LLM Orchestration SQLite3 (WASM) Docker & CI/CD 400,000+ Enterprise Scale @ Cognizant ASP.NET Core Angular PostgreSQL Redis Zero-Defect Production Releases AWS GenAI Practitioner 2026 Server-Side Prompt Injection Defense
Production Architecture HUD

Cognizant Token Defense Engine

Interactive security & scale simulator: Test how migrating validation logic from LLM prompts to server-side ASP.NET Core prevents prompt injection and cuts token costs by 90%.

1. Incoming Payload 3,200 Raw Tokens
➔
2. Server AST Guard PASSED (Sub-2ms)
➔
3. Model Context 320 Tokens (Top-K)
Server-Side Token Reduction: 90% Overhead Eliminated (Saving ~129.6M Tokens/Mo)
Simulated Daily Query Throughput: 48,000 queries/day
Open-Source Developer Tooling

VS Code Learning Agent

React 18 + TypeScript Extension Host with embedded SQLite3 (WASM) and Socratic hint streaming.

import { SocraticEngine, WASMDatabase } from '@learning-agent/core';

// 3-Tier Multi-LLM Orchestration + Embedded SQLite WASM
export async function streamGuidedHint(ctx: CodeContext) {
  const db = await WASMDatabase.init(); // < 10ms local query
  const history = await db.queryUserPath(ctx.userId);
  return SocraticEngine.generateHintStream({ history });
}
Technical Stack

Core Skills & Technologies

Hover card to organize
TypeScript
Next.js 15
React 18
FastAPI
Python
ASP.NET Core
Angular
PostgreSQL
SQLite3 WASM
Docker
Redis Caching
AWS GenAI
Agentic AI
Prompt Defense
GitHub Actions
Tailwind CSS
Enterprise Full-Stack Cloud

Salary Management Platform

Next.js 15 App Router + FastAPI managing 10,000+ seeded records with sub-15ms filtered search, Docker CI/CD, and Role-Based Access Control.

Scope Granted FULL_PAYROLL_WRITE
➔
Query Speed < 13.8ms (Indexed)
➔
Redaction Status 0% Masked (Audit Logged)
Explore Live Vercel Deployment ↗
Sai Jyothi Kumar Surla
Engineering Craft & Background

Sai Jyothi Kumar Surla

Associate Software Engineer @ Cognizant • Full-Stack & AI Systems
"In modern software engineering, our focus must be on security, end-user experience, and scalability. AI is an effective accelerator for research and automation, but production reliability still comes down to disciplined architecture and solid fundamentals."

Full-Stack Software Engineer with 3+ years of experience architecting distributed platforms and AI systems powering 400,000+ enterprise users at Cognizant. Experienced in scaling query throughput from 12k to 48k daily requests, building offline WASM diagnostic engines, and maintaining zero UI defect escalations across 8 production releases.

Direct Outreach

Let's Build Something Exceptional

Whether you have an ambitious full-stack engineering challenge, an AI agent system to build, or a key role to discuss, my inbox is open.