MOHIT RAJ DEO
"Architecting resilient systems that scale, self-heal, and drive impact."
The Person Behind the Systems
I'm Mohit Raj Deo — an engineer obsessed with building systems that are resilient, scalable, and self-healing.
I don't just write code — I architect infrastructure. From building distributed platforms that cut latency by 40%, to engineering backend systems that reduce cloud costs by 60%, I focus on systems that solve real problems at real scale.
< capabilities />
Distributed Systems
I design and build systems that handle thousands of concurrent operations — from distributed code execution engines to real-time streaming infrastructure. I think in terms of fault tolerance, horizontal scaling, and job orchestration.
Cloud Architecture & DevOps
I architect production infrastructure on AWS — auto-scaling groups, load balancers, container orchestration with EKS, and fully automated CI/CD pipelines. Zero-downtime deployments are my standard, not my goal.
Backend Engineering
I build backend platforms from scratch — REST APIs, database architecture, storage abstraction layers, and authentication systems. I've replicated core Supabase functionality with custom infrastructure.
AI & Machine Learning
From anomaly detection systems to reinforcement learning for healthcare optimization — I apply ML to solve real problems. My AI-powered monitoring system is a published patent.
< operations_log />
Eduniketan Private Limited (TheEduCode)
Distributed Code Execution Platform
/ Problem: Single-server execution couldn't handle concurrent users — high latency.
/ Solution: Built a distributed system using Judge0 CE, Aurora DB, Redis, and EC2 Auto Scaling.
Self-Hosted Backend Platform
/ Problem: Managed Supabase/S3 costs were high and created vendor lock-in.
/ Solution: Replicated core functionality with Backblaze B2 and SSH tunneling for secure access.
< research_lab />
AI-Powered Real-Time System Monitoring Platform
Addresses real-time monitoring challenges using Isolation Forest and KNN models to proactively detect anomalies and memory leaks, preventing an estimated 15-20% compute waste.
Multi-Objective Polypharmacy Optimization
Balancing therapeutic efficacy against ADR risk and tolerability using Pareto-optimal treatments across MIMIC-III and DDInter datasets.