Skip to content

(HELLO! I'M SUMANTH)

I BUILD SCALABLE BACKEND SYSTEMS

SUMANTH HS

(PORTFOLIO)

SOFTWARE ENGINEER

BACKEND | PLATFORM | DISTRIBUTED SYSTEMS

Engineering resilient backend platforms for high-throughput financial products.

I'm a creative engineer building production systems with a strong focus on delightful product performance and reliable digital infrastructure.

I currently work as a Software Engineer at CoinDCX, building Futures platform services that require strict data consistency and real-time synchronization. Previously at KreditBee, I built underwriting and OCR systems that improved disbursal capacity, reduced approval times, and lowered manual verification overhead.

PythonGo (Golang)RubyC++PostgreSQLSQLHTMLCSSFastAPIDjangoRuby on RailsFlaskAWS EC2AWS S3AWS LambdaAWS ECRKubernetesDockerApache KafkaRedisGitAsyncIOREST API DesignPythonGo (Golang)RubyC++PostgreSQLSQLHTMLCSSFastAPIDjangoRuby on RailsFlaskAWS EC2AWS S3AWS LambdaAWS ECRKubernetesDockerApache KafkaRedisGitAsyncIOREST API Design

3+

Years Experience

5+

APIs Delivered

120 Cr

Monthly Loan Throughput

40%

Latency Improvement

STACK

Skills

languages

PythonGo (Golang)RubyC++PostgreSQLSQLHTMLCSS

frameworks

FastAPIDjangoRuby on RailsFlask

infrastructure

AWS EC2AWS S3AWS LambdaAWS ECRKubernetesDocker

data And Ops

Apache KafkaRedisGitAsyncIOREST API Design

JOURNEY

Experience

Futures Platform

Software Engineer (Full-time)

CoinDCX

June 2025 - Present

Onsite

Built mission-critical backend for the futures domain with real-time synchronization between monolith and dedicated futures databases.

  • Designed and implemented a robust CDC pipeline with monitoring, retry workflows, and cron-based backfill to recover missed events.
  • Optimized order-placement APIs by replacing multiple DB lookups with a single Redis Lua flow, reducing request latency under load.
  • Improved operational confidence through consistency-first patterns and tighter failure recovery paths.
RubyRailsGoKubernetesAWSPostgreSQLRedis

Business Loan Rule Engine

Software Engineer (Full-time)

KreditBee

June 2023 - June 2025

Onsite

Led backend delivery for underwriting systems and external integrations powering business-loan workflows.

  • Built a microservice-based underwriting rule engine and 5+ REST APIs integrating vendors including TransUnion, PayU, and MobiKwik.
  • Reduced API response time from 5s to 3s (40% improvement) using asynchronous execution patterns.
  • Increased loan disbursal from 60 Cr to 120 Cr per month and reduced approval time by 17%.
  • Shipped an OCR pipeline using Google Cloud Vision to parse unstructured business documents and reduce manual effort by 15%.
PythonDjangoFastAPIAWSRedisMySQLGit

Bank Approval Rule Engine

Software Developer Intern

KreditBee

Feb 2023 - June 2023

Onsite

Contributed backend architecture for account-approval automation and fraud risk checks during loan disbursal.

  • Implemented fraud decision logic over tampering indicators, anti-money laundering signals, and blacklist sources.
  • Enabled automation of 350+ account approvals daily while reducing manual intervention.
  • Identified 30+ blacklist profiles per day, improving loan default control and reducing lending risk.
PythonRule EnginesRisk SystemsAutomation

BUILDS

Projects

NLP + Web Scraping

Reviews AZ

Built a sentiment analysis system for Flipkart product reviews using tokenization, Bag of Words features, and supervised classifiers.

Processed 10,000+ reviews with 93% prediction accuracy.

PythonDjangoBeautifulSoupNLPClassification

Computer Vision + Deep Learning

Anti-Spoofing Face Recognition

Developed an anti-spoofing pipeline to detect replay attacks from 2D image/video inputs using facial landmarks and behavioral prompts.

Improved spoof detection robustness with 3D facial landmark mapping and random-instruction checks.

PythonTensorFlowMediaPipeMTCNNDeep Learning

Recommendation System

WhichBook

Designed a recommendation engine using popularity and collaborative filtering over Kaggle book data with a Flask REST API.

Delivered a full-stack recommendation workflow with an API-backed frontend for interactive suggestions.

PythonFlaskRecommendation SystemsHTMLCSS

That's all, folks! Intrigued?

Let's Connect

Warning: Building products with me may result in an overload of growth, cleaner architecture, and happy users.

DIRECT LINKS

SEND A MESSAGE

I usually reply within 24 hours.