// platform engineer & Builder

Rakshit Jain.

Building systems that don't wake you at 3am.

I architect production-grade cloud systems serving millions — and build AI-native SaaS products focused on speed, scale, and intelligent automation.

View my work ↓GitHub ↗Resume ↓
Platform Engineer @ Quantiphi  ·  Builder, ScrollOS & AskMitra  ·  Open to interesting problems

10M+

Annual users served

50%

Latency reduction

3

Live SaaS products

99%

Uptime SLA

// 02 — experience

Where I've Worked

Platform Engineer

Quantiphi

Aug 2024 — Present

Mumbai, India

Designing and operating production infrastructure for global-scale systems.

  • ›Architected a real-time safety video analytics pipeline serving 10M+ annual visitors across 5+ global attractions with 99% uptime SLA.
  • ›Reduced production latency 50% (800ms → 400ms) via optimized data structures (hash maps, priority queues) and Kubernetes resource tuning.
  • ›Built high-throughput Python asyncio microservices processing 60+ frames/sec per stream with zero downtime and automatic retry.
  • ›Orchestrated CI/CD via GitLab + Helm, cutting deployment time 40% and improving team ramp-up time by 50%.
PythonKubernetesRedisGitLab CI/CDasyncio

Framework Engineer Intern

Quantiphi

Feb 2024 — Aug 2024

Mumbai, India

Designed reusable cloud-native infrastructure templates and optimized data pipelines.

  • ›Designed cloud-native application templates across AWS & GCP using Docker and Kubernetes, reducing new service deployment time by 92%.
  • ›Optimized automated ELT pipelines with Apache Airflow, improving data processing efficiency by 25% through parallelization and SQL query optimization.
AWSGCPDockerKubernetesApache Airflow

// 03 — selected shipments

Things I've Shipped

AskMitra

Founder & Lead Engineer

AI sales assistant for e-commerce

Live

An AI-powered sales assistant for e-commerce storefronts. Focused on revenue-driving automation — not generic chatbots.

  • ›Architected full AI pipeline: catalog ingestion → embedding → retrieval → conversational response.
  • ›Designed low-latency product recommendation engine for conversion-focused interactions.
  • ›Built merchant dashboard for monitoring AI performance and tuning responses.
  • ›Implemented structured LLM outputs with schema validation for production reliability.
Next.jsOpenAIGeminiPostgreSQLVector DB

ScrollOS

Founder & Lead Engineer

Intelligent email newsletter OS

Live

An end-to-end SaaS platform that intelligently organizes and classifies high-volume email newsletters. Not a CRUD app — a high-performance distributed system wrapped in clean UI.

  • ›Integrated Gmail/Outlook APIs via event-driven Node.js architecture; exponential backoff logic reduced inbox sync time 45% (10s → 5.5s).
  • ›Built custom ML-inspired classification engine (n-grams + Bayesian probability) achieving 90% accuracy across 1,000+ newsletters per user.
  • ›Improved system throughput 40% using Redis Pub/Sub, WebSockets, and batched DB writes (100+ records per flush).
  • ›PostgreSQL tsvector full-text search with sub-200ms response times at scale.
Next.js 15Node.jsPostgreSQLRedisWebSockets

TalentSync

Creator

AI-powered resume-job fit optimizer

Beta

An AI-powered resume-job fit optimization engine. Engineered for reliability — not demo-ware.

  • ›Designed multi-model fallback system (GPT-4 primary → GPT-4o-mini fallback) ensuring resilience during rate limits and outages.
  • ›Enforced strict type safety using Zod schema validation + OpenAI Structured Outputs, eliminating JSON parsing failures.
  • ›Built intelligent resume-job matching logic optimized for ATS systems.
Next.js 15TypeScriptPostgreSQLAWSOpenAI

// github activity

loading...

View profile ↗

loading contributions...

// 04 — technical arsenal

Tools of the Trade

Infrastructure & Cloud

AWS (EC2, Lambda, S3, RDS)GCPDockerKubernetesHelmGitLab CI/CDDistributed Systems Design

Backend

Node.jsExpress.jsFastAPIMicroservicesREST APIsEvent-driven architecture

Databases & Caching

PostgreSQLRedisMongoDBDynamoDBMySQL

Frontend

ReactNext.js 15TypeScriptTailwind CSSWebSocketsPerformance Optimization

Languages

TypeScriptPythonJavaScriptC++SQLBash

AI & Data

LLM API IntegrationPrompt EngineeringVector SearchStructured Outputsn-gram & Bayesian Modeling

Core Foundations

Data Structures & AlgorithmsSystem DesignLatency Optimization

// 06 — writing

Notes & Writing

50% Latency in 30 Days: What Hash Maps and Priority Queues Did to Our Video Pipeline

A deep-dive into how we reduced production video processing latency from 800ms to 400ms without touching the core algorithm. The fix wasn't clever — it was obvious in hindsight. Most performance wins are.

PythonasyncioPerformanceData Structures
12 min readFeb 2025↗

I Built a Bayesian Email Classifier Without a Single ML Library

ScrollOS classifies 1,000+ newsletters per user at 90% accuracy using n-gram frequency tables and Bayes' theorem — implemented from scratch in ~200 lines of TypeScript. Here's exactly how it works.

TypeScriptBayesianNLPSaaS
14 min readJan 2025↗

Kubernetes Resource Requests vs Limits: The Setting Nobody Explains Correctly

After tuning CPU/memory for real-time frame processing at 60fps, I have strong opinions on requests vs limits. The docs tell you what they do — they don't tell you what happens when you get it wrong at 3am.

KubernetesDevOpsPerformanceInfrastructure
11 min readDec 2024↗

GPT-4 Fallback Chains: Keeping Your AI Feature Alive When the Provider Isn't

TalentSync uses a GPT-4 → GPT-4o-mini fallback with Zod-validated structured outputs. The pattern is simple but the failure modes aren't. A practical guide to LLM reliability in production.

LLMsTypeScriptZodOpenAI
10 min readNov 2024↗

Redis Pub/Sub vs WebSockets: What I Learned Building Real-Time SaaS

ScrollOS uses both — and they solve different problems. After shipping 40% throughput improvements by combining them with batched writes, here's the mental model I use to decide which tool to reach for.

RedisWebSocketsNode.jsArchitecture
13 min readOct 2024↗

// 05 — behind the code

The Philosophy

I studied Electronics & Telecom — not Computer Science. Everything I know about systems, I built from first principles. That turned out to be an advantage. I don't reach for abstractions I don't understand. I read source code before documentation.

By day I'm a Platform Engineer at Quantiphi — building infrastructure that handles millions of users with strict SLAs. By night I ship AI-native SaaS products. The gap between "it works on my machine" and "it works for 10M people" is where I live.

Most performance problems are data structure problems. Most reliability problems are assumptions that weren't written down. I try to be explicit about both.

Based in Indore, India. Building for the internet.

Education

B.Tech — Electronics & Telecom

SGSITS, Indore

Graduated May 2024

Certifications

Associate Cloud Engineer

Google Cloud

Certified

Solutions Architect

AWS

Pursuing

Currently

  • ›Platform Engineer @ Quantiphi
  • ›Building ScrollOS & AskMitra
  • ›TalentSync in beta
  • ›Open to interesting problems

// 06 — get in touch

Let's Build.

If you're working on infrastructure problems, developer tools, or have something interesting to discuss — I'm all ears. Open to technical collaborations and product conversations.

Emailitsrakshitsjain@gmail.com
↗
GitHubgithub.com/rakshittt
↗
LinkedInrakshitbuilds
↗
X / Twitter@RakshitBuilds
↗

Built with Next.js 15 · Framer Motion · Deployed on Vercel