RM
RajatOS
Distributed SystemsCloud InfrastructureAI-Powered Systems

Building Systems.
Designed to Scale.

From distributed systems to intelligent applications, I build software where reliability, performance, and simplicity come first.

LEETCODE GUARDIAN
2,279
Top 0.6% Globally (Peak 2,379)
SCREENING SPEED
-75%
SignalForge Voice AI
ACCESSIBILITY
25+ Cmds
AuraMail Whisper STT
RESEARCH
IEEE '26
Paper Accepted
Rajat MalikSoftware Engineer
OFFICIAL CURRICULUM VITAE & SPECIFICATION

Resume & Track Record

Complete overview of software engineering accomplishments, research publications, and core technical proficiencies.

Rajat Malik

Verified RajatOS Specification

Software Engineer

Download PDF

Software Engineer specializing in backend engineering, distributed systems, and real-time AI infrastructure. Experienced building scalable RESTful APIs with Node.js, optimizing PostgreSQL databases, and developing production-ready applications supporting 10,000+ daily active users during a Software Developer Internship at Micro1. Architected SignalForge, a multi-tenant AI technical interviewing platform leveraging WebRTC, OpenAI Realtime API, Deepgram, and PostgreSQL, alongside AuraMail, an accessibility-first voice email system integrating Whisper, Gmail APIs, and ARIA-compliant interfaces. Passionate about system design, cloud-native software, and solving complex engineering problems through scalable architecture.

Primary Technical Stack
C/C++PythonSQLJavaScriptTypeScriptReactNext.jsNode.jsExpressFlaskPrisma ORMWebRTCTailwind CSS
Achievements & Publications
🏆 IEEE AMSCEHT 2026: Research paper accepted for presentation on an AI-powered voice-enabled email system.
🏆 Guardian (2279): Top 0.6% on LeetCode with 400+ algorithmic problems solved.
🏆 J.P. Morgan Code for Good: Runner-Up among 500+ participating teams for building a production-ready software solution under hackathon constraints.
🏆 APS HackFest: Winner for designing and delivering a functional software prototype within a 36–40 hour engineering sprint.
PDF format includes full career history, publications, & architecture diagrams
ENGINEERING CAPABILITIES & SPECIFICATION

Technical Skills

Core technologies, languages, distributed systems concepts, and Voice AI frameworks mastered across engineering projects.

29 Core Competencies

Languages

4 items
C++PythonJavaSQL

Systems programming, algorithm scripting, and database querying

Frameworks

5 items
ReactNext.jsNode.jsExpress.jsFlask

Modern frontend single-page & SSR runtimes with robust backend microservices

Databases & ORMs

4 items
PostgreSQLMySQLMongoDBPrisma ORM

Relational & document stores with typed ORM abstractions

Developer Tools

3 items
GitDockerPostman

Containerization, version control, and API testing workflows

Core CS

6 items
Data Structures & AlgorithmsSystem DesignOOPDBMSOperating SystemsComputer Networks

Algorithmic foundation, object orientation & high-scale distributed systems

AI & APIs

7 items
OpenAI APIDeepgramWhisperGemini AIREST APIsOAuthARIA

Speech recognition, generative AI models, secure auth & accessibility standards

FLAGSHIP PROJECT 1 OF 22026Voice AI & WebRTC Runtimes

SignalForge

Multi-Tenant AI Interviewing Platform

Architected a multi-tenant AI interviewing platform with 3 user roles, secure RBAC, WebRTC voice streaming, and automated candidate evaluation.

Recruiter OverheadAutomated
-40%
Manual scheduling reduction
Voice LatencyReal-Time
<500ms
WebRTC + OpenAI Realtime API
STT AccuracyDeepgram
95%+
Live bidirectional audio
Candidate ReviewReact Dash
30% Faster
PostgreSQL & Prisma schema
ARCHITECTURAL OVERVIEW & MOTIVATION

SignalForge is a multi-tenant AI interviewing platform built with React, Express.js, PostgreSQL, Prisma, OpenAI Realtime API, Deepgram, and WebRTC. It automates first-round technical screening supporting 3 user roles with secure RBAC, conducts low-latency (<500ms) voice interviews with 95%+ transcription accuracy, and provides recruiters with dynamic dashboards and automated candidate evaluations.

TECH STACK:ReactExpress.jsPostgreSQLPrismaOpenAI Realtime APIDeepgramWebRTCTailwind CSS
INTERACTIVE SYSTEM ARCHITECTURE TOPOLOGY

SignalForge Pipeline Graph

Click any node below to inspect execution flow, latency budget, and low-level subsystem implementation details.

Real-Time Stream Pipeline Active

WebRTC Voice Gateway

Real-time Audio Stream Pipeline

LATENCY<50 ms

Deepgram & Whisper STT/TTS

Speech-to-Text Processing

LATENCY~250 ms

OpenAI & Gemini AI Evaluator

Automated Technical Scoring

LATENCY~600 ms

PostgreSQL & RBAC Auth Engine

Multi-Tenant Workflows

LATENCY<5 ms
SUBSYSTEM INSPECTORGATEWAY

WebRTC Voice Gateway

Real-time Audio Stream Pipeline

Low-latency WebRTC peer connections routing concurrent voice input directly to processing workers.

Execution Stage:WebRTC MediaStream Track
Allocations / Call:V8 Garbage Collection
Target Budget:<50 ms
RMRajatOS Architecture Spec
ENGINEERING TRADE-OFFS & DECISIONS

Why Options Were Chosen vs. Rejected

OpenAI Realtime API + Gemini AI over static text prompts

Static text models introduced 3+ second latencies unsuited for interactive voice conversations.

CHOSEN APPROACH:Streaming WebRTC audio directly to Deepgram STT and processing evaluations asynchronously.
ALTERNATIVE REJECTED:Synchronous REST endpoint polling per spoken sentence.
BENEFIT IMPACT:Reduced screening turnaround times by 75% while producing instant candidate scorecards.
View Source CodeLive Production Specs
FLAGSHIP PROJECT 2 OF 22025Voice AI & ARIA Accessibility

Voice-Based Email System

Hands-Free Voice-Powered Email Platform

Built a hands-free, AI-powered email platform supporting 25+ concurrent voice commands with Whisper STT, gTTS, and Gmail API.

Voice Commands100% Hands-Free
25+
Compose, search, & inbox management
STT AccuracyWhisper
98%
Flask backend orchestration
AccessibilityWCAG AAA
ARIA Strict
Designed for visually impaired users
Task SpeedNext.js
25% Faster
Real-time voice feedback
ARCHITECTURAL OVERVIEW & MOTIVATION

Voice-Based Email System (AuraMail) is a hands-free, AI-powered email client designed to enable visually impaired users to compose, search, and manage emails using 25+ voice commands. Built with React, Next.js, Flask, Whisper Speech-to-Text (98% accuracy), gTTS, and Gmail API adhering strictly to ARIA accessibility standards.

TECH STACK:ReactNext.jsFlaskWhispergTTSGmail APITailwind CSSARIA
INTERACTIVE SYSTEM ARCHITECTURE TOPOLOGY

Voice-Based Email System Pipeline Graph

Click any node below to inspect execution flow, latency budget, and low-level subsystem implementation details.

Real-Time Stream Pipeline Active

Whisper Speech-to-Text Engine

Voice Command Parser

LATENCY~250 ms

Gmail API Integration Handoff

Email Dispatch & Sync

LATENCY~150 ms

ARIA Screen Reader Engine

Tactile Accessibility Layer

LATENCY<10 ms
SUBSYSTEM INSPECTORCOMPUTE

Whisper Speech-to-Text Engine

Voice Command Parser

Local & cloud Whisper speech recognition achieving 98% accuracy mapping spoken intent to 25+ email actions.

Execution Stage:Node.js Stream Piping
Allocations / Call:V8 Garbage Collection
Target Budget:~250 ms
RMRajatOS Architecture Spec
ENGINEERING TRADE-OFFS & DECISIONS

Why Options Were Chosen vs. Rejected

Whisper STT with client-side ARIA focus management over native browser web speech

Browser WebSpeech API lacked consistency across screen readers and offline voice accuracy.

CHOSEN APPROACH:Fused Whisper STT with custom gTTS audio playback and ARIA state announcements.
ALTERNATIVE REJECTED:Default WebSpeech API with standard HTML buttons.
BENEFIT IMPACT:Achieved 98% transcription accuracy and 100% hands-free email composition for visually impaired users.
View Source CodeLive Production Specs
TECHNICAL PUBLICATIONS & DEEP DIVES

Engineering Writings

In-depth architectural analysis, real-time voice engineering, accessibility standards, and algorithms.

Software Philosophy6 min read

What Hackathons Taught Me About Shipping Software

For a long time, I believed software engineering was about writing good code. Hackathons taught me that good code is only a small part of the job.

OCT 2025Read Article
Voice AI & WebRTC6 min read

Architecting Real-Time Voice Screening with WebRTC & OpenAI Realtime API

Building SignalForge required converting traditional REST API polling into continuous WebRTC voice streams integrated with Deepgram STT and Gemini AI evaluations.

MAY 2026Read Article
Accessibility & AI7 min read

Designing Hands-Free Web Applications: ARIA Accessibility & Whisper STT

Combining Whisper Speech-to-Text and gTTS with comprehensive ARIA accessibility tags allowed us to build AuraMail, an accessible voice-driven email experience.

DEC 2025Read Article
Algorithms & CS5 min read

Mastering Competitive Programming: Strategies Behind a 2,279 LeetCode Guardian Rating

Reaching Guardian level (2,279 rating, peak 2,379) builds deep intuition for memory bounds, time complexity, and low-latency data structures in production applications.

OCT 2025Read Article
ENGINEERING CRAFTSMANSHIP & PRINCIPLES

Engineering Philosophy

Notes from the journal of building resilient distributed systems, core backend infrastructure, and software designed to scale.

Simplicity is a Hard-Won Choice

Clever code is a liability; clear code is a force multiplier.

Latency is a Budget, Not a Metric

A system that is occasionally fast is effectively slow.

Failure is the Normal State

Assume every dependency, network, and node will fail simultaneously.

State is the Root of All Evil

Isolate state mutation, keep compute pure, and push side-effects to the edge.

Observability Trumps Guesses

If you cannot measure internal state, you are operating on hope.

CHALKBOARD AXIOM
tenet-1
Core Belief:

The hardest part of software engineering isn't solving complex problems—it's preventing complex solutions.

Rule of Thumb:

If you can't debug it in your head during an outage, you haven't finished designing it.

ENGINEERING JOURNAL NOTEBOOK

Simplicity is a Hard-Won Choice

Clever code is a liability; clear code is a force multiplier.

In production, cleverness degrades faster than hardware. Systems rarely break because an algorithm wasn't sophisticated enough; they break because someone couldn't trace state under pressure at 3 AM. Every additional layer of abstraction is a promise you have to keep forever.

I optimize for legibility over brevity. If an engineer cannot trace a data flow in five minutes without an architecture diagram, the implementation is wrong.

Engineering Practices:
Write explicit data transformations instead of magic runtime behaviors.
Limit indirection to boundaries where implementations actually change.
Delete unused abstraction layers before adding new features.
Treat code removal as a higher-value output than code creation.
Real Engineering Trade-Off:

Chose a cohesive monolithic domain layout over distributed micro-services during early service exploration—trading hypothetical independent deployment speed for unified transactional boundaries, instant local debugging, and zero network call latency.

Implementation Pattern:
// Explicit, transparent data flow without hidden framework magic
async function processExecutionPayload(
  payload: IncomingRequest
): Promise<ExecutionResult> {
  const sanitized = sanitizeRequest(payload);
  const authContext = await authenticateSession(sanitized.headers);
  const record = await db.transactions.insert({
    userId: authContext.id,
    data: sanitized.body
  });
  return { success: true, transactionId: record.id };
}
RM Engineering JournalPRODUCTION TESTED
RAJATOS INTERACTIVE SHELL

Developer Terminal

Execute direct commands to query low-level system telemetry, architecture specs, and engineer profile facts.

zsh — rajat@apple-m3-max: ~
UTF-8
rajat@m3-max:~$whoami --verbose

NAME: Rajat Malik

ROLE: Software Engineer

FOCUS: Distributed Systems • Backend Engineering • Cloud Infrastructure • AI Systems

LOCATION: India (Open to Global Opportunities)

STATUS: Open to exceptional engineering opportunities

rajat@m3-max:~$sysctl -a engineering.profile

ENGINEERING INTERESTS:

• Distributed Systems

• Backend Infrastructure

• Cloud Platforms

• AI-Powered Applications

• System Design

COMPETITIVE PROGRAMMING:

• LeetCode Guardian (Peak 2379)

• Top 0.6% Global

• 400+ Problems Solved

ENGINEERING PHILOSOPHY:

Build systems that remain simple,

reliable, and scalable as they grow.

Type 'help' to view all interactive terminal shell commands.

rajat@m3-max:~$
CAREER & TRACK RECORD

Engineering Experience

Building scalable backend systems, optimizing production infrastructure, and shipping reliable software in collaborative engineering teams.

Jan 2026 — May 2026

Software Developer Intern

Micro1

Remote

Built production RESTful APIs with Node.js, optimized PostgreSQL query performance supporting 10,000+ daily active users, developed reusable React components, and collaborated within a distributed Agile engineering team to deliver scalable software features.

KEY ENGINEERING CONTRIBUTIONS:
Engineered RESTful APIs using Node.js and optimized complex PostgreSQL queries, reducing data retrieval latency by 30% and supporting 10,000+ daily active users.
Developed 15+ reusable, responsive frontend components in React, increasing code maintainability and decreasing UI development time by 20%.
Collaborated within a distributed Agile team of 8 engineers to deliver modular features via Git version control and code reviews, reducing post-deployment bugs by 25%.
Node.jsPostgreSQLReactGitREST APIsAgile
CONTACT
DIRECT COMMUNICATION PORTALResponse SLA < 2 Hours

One More Thing... Let’s Connect.

Direct channel for core engineering lead roles, distributed systems advisory, or technical keynote inquiries.

DIRECT MESSAGE DISPATCHChannel Ready
PRIMARY CONTACT24/7 Monitored

rajatmalik.dev@gmail.com

Whether it’s an engineering opportunity, collaboration, or simply a technical conversation, I’d love to hear from you.

OPERATIONS BASEIST (UTC+5:30) ACTIVE

Based in India • Available Globally

Open to Software Engineering, Backend, Full-Stack, and AI Infrastructure opportunities.

VERIFIED PROFILES