BACKEND & SYSTEMS ENGINEER
BUILDER · DEVELOPER · ALWAYS LEARNING
Building
software that actually
works.
I'm Arjun — a backend engineer and builder currently interning at WhatBytes. I engineer high-throughput scraping pipelines, distributed task queues with Celery & Redis, and low-latency voice RAG.
FOCUS
Backend & Distributed Queues
ACADEMICS
B.Tech CSE · CGPA 9.5
SYSTEM ARCHITECTURE
“Distributed queues, low latency, calm software.”
LearnLoop
AI workspace for real learning and retention.
RoBERTa Legal AI
Agentic legal document analysis & contract risk assessment.
Vaani
Voice-enabled RAG system for real conversations in Hindi.
HireLens AI
AI-powered engineering portfolio & code translation.
Episodic Memory Platform
Time-aware semantic memory & episodic RAG architecture.
Lynx
Grounded Linux system action engine & voice assistant.
Nook 🌱
Local-first cozy browser companion & tab terrarium.
Notion Job Tracker
One-click job application ingestion directly to Notion.
BACKEND PHILOSOPHY
“Good backend architecture doesn't shout. It handles thousands of asynchronous jobs quietly and fails gracefully.”
Where I've built & delivered.
Engineering production backends, asynchronous worker pipelines, and scalable APIs.
WhatBytes
InternshipFull Stack Developer Intern
Remote
Engineering asynchronous distributed scraping pipelines, real-time community chat architectures, and sub-40ms feed-ranking engines under high concurrency.
98%
Ban Rate Cut
65%
Latency Cut
< 40ms
Feed Query
500+
Daily Ingestion
Architected an asynchronous distributed scraping pipeline across 30+ media platforms using Python, Playwright, Scrapy, and Celery, ingesting 500+ daily records with proxy rotation to cut ban rates by 98%.
Designed real-time community chat and media pipelines using Django Channels and Redis, cutting video and OCR processing latency by 65% via parallel Celery workers (FFmpeg/Tesseract).
Engineered a hybrid feed-ranking engine combining time-decay heuristics with zero-shot classification, keeping query execution under 40ms.
Implemented transactional workflow state machines, leveraging row locks and atomic rollbacks to guarantee data consistency under high concurrent loads.
30+ Media Ingestion Feeds & <40ms Feed Querying
Designed parallel Celery scraping tasks paired with dynamic IP proxy rotations, reducing ban rates by 98% while streaming real-time community chats via Django Channels and Redis pub/sub.
~70% Quotation Turnaround Reduction
Automated line-item extraction from 25+ procurement PDFs and scans using OCR + LLMs, automatically cross-referencing inventory across 20+ monthly orders with structured tracing.
Stealth Startup
InternshipSoftware Engineer Intern
Remote
Engineered an automated document intelligence extraction pipeline turning messy procurement orders and scans into verified inventory records.
~70%
Turnaround Cut
-60%
Debugging Time
20+
Reconciled Orders
Cut quotation turnaround time by ~70% by building an OCR + LLM pipeline to extract line items from 25+ procurement PDFs and images.
Improved quotation accuracy across 20+ monthly orders by matching extracted items against live inventory.
Reduced debugging time by ~60% by adding structured logging and error tracing across the OCR-to-quotation pipeline.
Bennett University
Bachelor of Technology in Computer Science
Top 5%
Rigorous computer science curriculum emphasizing algorithm analysis, distributed databases, machine learning systems, and operating systems.
HONOR ROLL
Dean's List for Academic Excellence
Open Source Systems
Architected and shipped production-ready open source systems including LearnLoop, RoBERTa Legal AI, and high-throughput vector platforms.
University Hackathons
Ranked in the Top 20 across multiple competitive hackathons; awarded Dean's List for sustained academic excellence.
Hyperspectral Research
Authored research on UAV hyperspectral imaging for wheat yield prediction utilizing the Minnesota DRUM agricultural dataset.
WhatBytes
InternshipFull Stack Developer Intern
July 2026 – Present · Remote
Engineering asynchronous distributed scraping pipelines, real-time community chat architectures, and sub-40ms feed-ranking engines under high concurrency.
98%
Ban Rate Cut
65%
Latency Cut
< 40ms
Feed Query
500+
Daily Ingestion
Architected an asynchronous distributed scraping pipeline across 30+ media platforms using Python, Playwright, Scrapy, and Celery, ingesting 500+ daily records with proxy rotation to cut ban rates by 98%.
Designed real-time community chat and media pipelines using Django Channels and Redis, cutting video and OCR processing latency by 65% via parallel Celery workers (FFmpeg/Tesseract).
Engineered a hybrid feed-ranking engine combining time-decay heuristics with zero-shot classification, keeping query execution under 40ms.
Implemented transactional workflow state machines, leveraging row locks and atomic rollbacks to guarantee data consistency under high concurrent loads.
Stealth Startup
InternshipSoftware Engineer Intern
December 2025 – April 2026 · Remote
Engineered an automated document intelligence extraction pipeline turning messy procurement orders and scans into verified inventory records.
~70%
Turnaround Cut
-60%
Debugging Time
20+
Reconciled Orders
Cut quotation turnaround time by ~70% by building an OCR + LLM pipeline to extract line items from 25+ procurement PDFs and images.
Improved quotation accuracy across 20+ monthly orders by matching extracted items against live inventory.
Reduced debugging time by ~60% by adding structured logging and error tracing across the OCR-to-quotation pipeline.
Bennett University
Bachelor of Technology in Computer Science
Open Source Systems
Architected and shipped production-ready open source systems including LearnLoop, RoBERTa Legal AI, and high-throughput vector platforms.
University Hackathons
Ranked in the Top 20 across multiple competitive hackathons; awarded Dean's List for sustained academic excellence.
Hyperspectral Research
Authored research on UAV hyperspectral imaging for wheat yield prediction utilizing the Minnesota DRUM agricultural dataset.
From ideas to real systems.
Swipe, click side cards, or use arrow keys to navigate the 3D project deck.
AI workspace for real learning and retention.
An intelligent workspace that transforms lectures, PDFs, and web articles into structured mental models, adaptive quizzes, and contextual notes with specialized cognitive modes.
Agentic legal document analysis & contract risk assessment.
A fine-tuned RoBERTa platform for legal document analysis, RAG-based retrieval with agent task decomposition, and clause risk classification across 300+ legal documents.
Voice-enabled RAG system for real conversations in Hindi.
A speech-to-speech conversational AI system built for Hackathon Goa 2026. Streams Hindi audio end-to-end using Sarvam AI STT/TTS, FAISS vector retrieval over MS MARCO-X!, Groq-accelerated generation, and 4 strict hallucination guardrails.
AI-powered engineering portfolio & code translation.
Transforms fragmented GitHub commit histories into cohesive, recruiter-ready profiles with architectural scoring, technical metrics, and 3D contribution graphs.
Time-aware semantic memory & episodic RAG architecture.
Production-style cognitive memory backend modeling user knowledge as time-stamped semantic memory events with Endee Vector DB (SIMD C++ / RoaringBitmaps), OpenRouter LLMs, and exponential time decay.
ACTIVE: Vaani · CLICK SIDES OR ARROWS TO ROTATE
Engineering with focus & craft.
Hey, I'm Arjun. I'm a software engineer and builder, currently working as a Full Stack / Backend Developer Intern at WhatBytes.
Most of my engineering revolves around backend systems, asynchronous task queues with Celery & Redis, web scraping pipelines, and domain-grounded retrieval (RAG). I like building tools that are fast, deterministic, and handle messy real-world data without breaking.
I maintain a 9.5 / 10 CGPA (Top 5% · Dean's List), and spend my free time exploring Linux system internals and shipping production-ready open-source projects like LearnLoop, RoBERTa Legal AI, and Vaani.
B.Tech in Computer Science
Bennett University · CGPA 9.5 (Dean's List)
Backend / Full Stack Intern
WhatBytes · Django REST, Celery, Redis
8+ Flagship Projects
Production systems, RAG, & vector tooling
Dean's List & Hackathons
Top 20 in university hackathons · ML Research
Arjun Chaudhary — Backend Engineer & Builder.
Currently a Backend / Full Stack Intern at WhatBytes, building asynchronous task queues, high-throughput scrapers, and content ranking logic.
Type help or click the chips below to explore commands.