> sanjay.init()
Sanjay Raju
Software Engineer & Founder
I take a business problem and own it end to end — product, architecture, code, deployment. Currently building a retail operations platform for high-SKU trade shops in the GCC, and taking on select builds: ERP and retail systems, speed-sensitive backends, AI platforms, and cloud infrastructure that scales.
what i build
I'm not an agency. I'm an engineer who builds products — and I take on select client work where that mindset pays off: one person who owns the outcome end to end, from requirements and architecture to deployment, training, and support.
Retail & ERP Systems
Point-of-sale, inventory, VAT-compliant invoicing, customer credit, repairs, and reporting — built for real shop floors. Offline-resilient systems that keep billing even when the internet drops.
High-Performance Backends
Custom, speed-sensitive systems where performance is the product — real-time data pipelines, low-latency APIs, GPU-accelerated processing, and architectures measured in milliseconds.
Custom Software
Web platforms, iOS apps, internal tools, and dashboards built end-to-end — and integrated with the systems your business already runs on.
AI, RAG & Automation
Custom RAG pipelines and AI search systems built over your company's own data, plus LLM-powered features, document processing, and workflow automation — production-grade, with guardrails and security built in.
Cloud & Infrastructure
Architecture that scales with your business — AWS deployments, CI/CD pipelines, migrations off legacy systems, monitoring, and keeping your cloud bill under control.
- ERP & retail management systems — POS, inventory, VAT invoicing
- Custom RAG pipelines & AI search over your company's data
- High-performance, speed-sensitive backend systems
- Web platforms, dashboards & internal tools
- iOS & mobile apps
- Business-process & workflow automation
- Cloud architecture & AWS deployments
- Legacy system replacement & data migration
- CI/CD, monitoring & ongoing support
about
I'm Sanjay Raju — an engineer and founder based in Bahrain. I hold a Master's in Data Science from Stevens Institute of Technology and a B.Tech in Artificial Intelligence from Amrita Vishwa Vidyapeetham, and I build complete systems: the product thinking, the architecture, the code, and the infrastructure it runs on.
Right now I'm building a retail operations platform for high-SKU trade and electronics shops in the GCC. Before that, I founded and built a Gen-AI powered news platform end-to-end — E2E encrypted messaging, a hybrid BM25 + vector search RAG, a SwiftUI iOS app, and a FastAPI backend on AWS ECS/Fargate — as a one-person engineering team.
My background is in performance-critical systems: at Tata Elxsi I built CUDA-accelerated C++ pipelines for video-processing ML models, and my deep learning research on lunar crater detection was published in the Journal of the Indian Society of Remote Sensing. If your problem needs software — from a shop-floor billing system to a speed-sensitive backend — I can build it.
building
Retail Operations Platform
The Operating System for GCC Trade Shops
A retail management platform for electrical and electronics merchants in the Gulf — multi-counter POS, 50,000+ SKU inventory, VAT-compliant invoicing, repair workflows, and a local-first architecture that keeps the shop billing through internet and LAN outages, with real-time cloud sync for owners. Launching with a 37-year-old trade business in Bahrain as the proving ground.
S2S MK1
Full-Duplex Indic Speech Translation
Designing the first real-time, full-duplex speech-to-speech translation system for Indian languages — preserving the speaker's voice and accent across Tamil, Telugu, Kannada, Malayalam, and Hindi. Novel architecture built on Moshi's inner monologue, extended with a three-stage per-frame translation chain.
projects
Personalized News Platform
Gen-AI Powered Personalized Short-Form Video News
A personalized news platform with custom political alignment algorithm, E2E encrypted messaging, and hybrid AI-powered search and retrieval RAG — built for iOS with a FastAPI backend on AWS.
- Custom political alignment algorithm for content personalization
- E2E encrypted messaging with Signal Protocol patterns
- Hybrid search & retrieval RAG: BM25 + pgvector HNSW + OpenSearch + Bedrock embeddings
- Multi-role Docker containers on AWS ECS/Fargate
- CI/CD with GitHub Actions and AWS ECR
- SwiftUI/UIKit iOS app with Supabase RealTime
Superhuman
A Tauri desktop app that captures meeting audio in real-time, transcribes speech locally with Whisper, and uses Claude to instantly detect and answer questions — keeping you sharp without breaking focus.
View DetailsFolioFin
A native Linux desktop reading client for Jellyfin media servers. Supports EPUB, PDF, and CBR comics with a dark reader UI, per-book progress sync to Jellyfin, and full library browsing.
View DetailsJob Aggregator AI Agent
An intelligent system that parses resumes, scrapes 1000+ job listings, and matches them using AI embeddings with a conversational interface.
View DetailsLunar Crater Detection
Deep learning model detecting craters on the lunar surface from Chandrayaan-2 satellite imagery. Published in the Journal of the Indian Society of Remote Sensing.
View DetailsImage Captioning System
Multi-model system using LSTM and Feed Forward Neural Networks on the Flickr30k dataset, achieving 78% accuracy in image recognition and captioning.
View DetailsMental Health in Tech Analysis
Deep analysis of survey data correlating workplace conditions and mental health in tech using statistical and Bayesian methods.
ML Video Processing Pipeline
Large-scale video processing pipeline using C++ for multiple ML models with CUDA acceleration, built with a team of 30 at Tata Elxsi.
Driver Drowsiness Detection
Real-time system using OpenCV for facial landmarks and a CNN model to detect drowsiness with early yawning warnings.
experience
Founder & Full-Stack Engineer
Personalized News Platform (Independent) · 2025 - Present
- Designed and built a Gen-AI powered personalized news platform end-to-end
- FastAPI backend with multi-role Docker containers on AWS ECS/Fargate, ALB, CloudFront CDN
- Implemented E2E encrypted messaging using Signal Protocol patterns (X25519, AES-256-GCM)
- Built hybrid search and retrieval RAG: PostgreSQL BM25 + pgvector HNSW + AWS Bedrock Titan embeddings + OpenSearch with reciprocal rank fusion
- Developed iOS app in SwiftUI/UIKit with Supabase RealTime for instant messaging
- CI/CD pipelines with GitHub Actions and AWS ECR
AI Engineer Intern
Tata Elxsi · Jan 2023 - Jul 2023
- Worked with a team of 30 on a multimillion-dollar video processing and interpretation project
- Created a unified C++ pipeline for processing multiple ML models
- Enhanced performance by implementing CUDA acceleration
- Built multithreaded architecture for simultaneous input processing
Full Stack Developer
Al Qayam Electrical Equipments W.L.L. · 2024
- Built website for a local business to market internationally in the Middle East
- Designed pipelines to direct product enquiries and orders to company software
education
Master's in Data Science
Stevens Institute of Technology
Aug 2023 - May 2025
B.Tech in Artificial Intelligence
Amrita Vishwa Vidyapeetham
2019 - 2023 · GPA: 8.8/10.0
blog
Building a Hybrid Search Engine with BM25, Vector Embeddings, and Real-Time SSE Streaming
Search in a video-first content platform is fundamentally different from traditional web search. Users type "Israel ceasefire" and expect not just matching videos, but an AI-generated news briefing, relevant creators, trending topics, and all of it appearing progressively - creators first (100ms), then videos in chunks, then an AI summary streaming token-by-token like a ChatGPT response. A single monolithic JSON response would mean 3-5 seconds of staring at a blank screen.
Read MoreThe 7,000-Line UICollectionViewCell: Building a TikTok-Style Video Feed That Doesn't Stutter
How I built a 4-tier video caching system and a single UICollectionViewCell that handles playback, engagement tracking, gesture recognition, trust visualization, recovery from compositor crashes, and first-frame detection for HLS streams - all at 60fps.
Read MoreHow I Built Signal-Level E2E Encrypted Messaging
Building a Signal-level encryption into a consumer app: A Technical deep dive
Read Morecontact
Running your business on paper, spreadsheets, or a legacy system nobody can maintain? Tell me how your business works — I'll tell you what's possible. First conversation is free, no obligation.
$ cat availability.txt
Open to new projects — ERP & retail systems, custom software, AI, cloud.
$ echo $EMAIL
$ echo $RESPONSE_TIME
Usually within a day.