Anurag Soni

Software Engineer

India2 yrs experience

Highly Stable

Key Highlights

Built scalable systems serving millions of users.
Led serverless migration reducing costs by 60%.
Developed multilingual systems for diverse user base.

Stackforce AI infers this person is a Backend-heavy Fullstack Engineer in the EdTech industry.

Contact

Skills

Core Skills

Node.jsAwsLlmsAlgorithms

Other Skills

React.jsAWS FargatepgCronFirehoseRedshiftAWS LambdaAWS RedshiftS3API load testingC++TypeScriptDatabasesAWS Elastic BeanstalkAmazon Relational Database Service (RDS)Amazon Dynamodb

About

I'm a software engineer who builds things that actually run at scale. At ConveGenius, I work on production infrastructure serving 1.5 crore students and 1 lakh teachers across 5 Indian states — pipelines, OCR systems, LLM workflows, serverless architecture. A few things I've shipped: → Data pipelines processing 100K–200K records/day (Firehose → Redshift) → OCR pipeline on AWS Lambda that processes ~1 crore answer sheets every 2 months → LLM-powered multilingual systems used by 50K+ users → Serverless migration that cut infra costs by 60% Outside of work, I'm building my own products. I'm documenting the whole journey publicly: what I build, what breaks, what I learn. I'm looking for my next role: backend or full-stack, global remote or relocation. Ideally a product company where scale and impact actually mean something. Tech: Node.js · TypeScript · React · AWS (Lambda, Fargate, S3, Redshift, Firehose) · PostgreSQL · Redis · LLMs · System Design Let's connect if any of this overlaps with what you're building.

Experience

2 yrs

Total Experience

2 yrs

Average Tenure

2 yrs

Current Experience

Convegenius.ai

2 roles

Software Engineer

Jun 2024 – Present · 2 yrs

Building production systems at government scale across 5+ Indian states.
→ Contributed to a MIS platform serving 1.5 crore students and 1 lakh teachers
→ Built Grade Progression workflows with async processing (AWS Fargate, pgCron)
for bulk academic operations affecting hundreds of thousands of students
→ Designed data pipelines processing 100K–200K records/day
(Firehose → Redshift, S3 ETL)
→ Engineered an OCR evaluation pipeline on AWS Lambda processing ~1 crore
answer sheets every 2 months — a core revenue-contributing product
→ Led serverless migration to AWS Lambda, reducing infrastructure costs by ~60%
→ Built LLM-powered multilingual systems (3+ languages) for 50K+ users
→ Designed multi-agent workflows for structured data extraction from APIs and PDFs
→ Built internal tooling that reduced developer effort by ~40% (~15 hrs/week saved)