Engineer focused on Backend, Data Engineering, and Scalable Systems Architecture. I build distributed systems, data pipelines, and cloud infrastructure that power real-world applications at scale. Off-keyboard you'll find me on Arch-based distros, playing chess, or cycling.

name: Suraj Maurya
title: Backend Engineering · Data Engineering · Scalable Systems
location: Mumbai, India
current_role: Technical Lead
previously: [Skima.ai, Shaadi.com]
focus:
  - Distributed Systems & High-Throughput Backends
  - Data Pipelines & ML Platforms
  - Cloud Infrastructure & Observability
languages: [Go, Python, C++, Bash]
hobbies: [Arch Linux, Open Source, Chess, Cycling]
ask_me_about: kafka, kubernetes, feature_stores, etl, scaling

End-to-end automated technical interviewer with real-time voice AI. Go (Gin) backend orchestrating Python-based LiveKit Agents, integrating Deepgram Nova-3 (STT) → GPT-4o (LLM) → ElevenLabs Turbo v2.5 (TTS) into a sub-second conversational pipeline. Built reconnection / state-sync handlers for graceful network interruption recovery, a two-stage LLM prefill that dynamically adapts role-specific question flows, async transcript streaming to MongoDB with exponential backoff, and RabbitMQ-driven post-interview scoring. Deployed on AWS EKS with rolling updates.

Go Gin Python asyncio LiveKit GPT-4o MongoDB RabbitMQ EKS

Horizontally scalable Kafka platform processing 100M+ events/day with exactly-once semantics, CQRS, and idempotent writes. At-least-once delivery, fault isolation, and regional failover for 3 business-critical services — sustained 99.99% uptime and reduced failover recovery time by 90%. Self-healing ETL pipelines with automated schema diffing and versioning support backward-compatible evolution; powered analytics across 5M+ daily records.

Go Kafka Kubernetes DynamoDB Datadog

Modular, versioned Feature Store (inspired by Feast) with offline/online sync — powering 30+ ML models at sub-150 ms latency via Redis caching with LRU eviction. Cut model deployment lag from 4 hours to under 5 minutes, halved model-onboarding time through a standardized schema catalog spanning MySQL, DynamoDB, and Kafka. Owned the low-latency feature delivery API (Go + AWS API Gateway) at 99.99% SLA; partnered with revenue and data-science teams to lift conversion by 25%. Companion Redshift snapshot orchestration service with rollback support enabled daily ML retraining and cut training backlog by 70%.

Go Redis Redshift MySQL DynamoDB Kafka AWS API Gateway

Fault-tolerant, multithreaded scraper in Go with rotating proxies, rate-limiting, and dynamic selectors — monitored 5+ competitor sites with zero-downtime config reloads. Cut competitive-insight turnaround time by 90%, helping the GTM team improve campaign ROI by 60%.

Go Concurrency Proxy Rotation

4+ years of distributed backend, data infrastructure, and platform engineering.

AWS depth: EC2 · ECS · EKS · Lambda · S3 · RDS · DynamoDB · ElastiCache · CloudWatch · Redshift · API Gateway

Concepts: SQL · ETL · Data Pipelines · Data Warehouse · Data Lake · Data Catalog · Schema Evolution · CQRS

Logging · Alerting · SLO/SLA Metrics · Anomaly Detection · MTTR Optimization

REST APIs · Microservices · Event-Driven Architecture

Bachelor of Engineering, Computer Science — Shree L.R. Tiwari College of Engineering, University of Mumbai · CGPA 9.03 / 10 · 2018 – 2022

• 🌟 2× "Rockstar of the Month" among 100+ engineers at Shaadi.com — for delivering scalable, production-grade systems with high impact
• 🛠️ Maintainer of 30+ internal backend repositories supporting core real-time services and APIs
• 🧑‍🏫 Mentored junior engineers to production-readiness in under 6 weeks; delivered internal tech talks on back-pressure & schema evolution to 40+ engineers
• 🥇 Winner, CODEYANTRA CODEY 2.0 national coding competition (2020)
• 🧠 300+ DSA problems solved across CodeChef, LeetCode, and GFG
• ⭐ CodeChef 3★ — peak rating 1603

_{"Build for scale. Fail fast. Ship reliably."}