Sumanth H S
Software Engineer

Contact
- Mobile: +91-9900325989
- Email: hsumanth2707@gmail.com
- LinkedIn: https://www.linkedin.com/in/sumanth27/
- GitHub: https://github.com/Sumanthhs27

Education
1. Indian Institute of Technology (IIT) Roorkee
   M.Tech - Data Science | GPA: 8.71
   Sept 2021 - Sept 2023 | Roorkee, India
   Coursework: Data Structures and Algorithms, Machine Learning, Deep Learning, Databases

2. BMS College of Engineering
   B.Tech - Mechanical Engineering | GPA: 8.76
   Sept 2016 - Sept 2020 | Bangalore, India

Skills
- Languages: Python, Go (Golang), Ruby, C++, PostgreSQL, SQL, HTML, CSS
- Frameworks: FastAPI, Django, Ruby on Rails, Flask
- Tools: AWS (EC2, S3, Lambda, ECR), Kubernetes, Docker, Apache Kafka, Redis, Git

Experience
1. CoinDCX | Software Engineer (Full-time) | June 2025 - Present | Onsite
   - Built a robust CDC pipeline between monolith and futures DB for real-time sync.
   - Added monitoring, retry, and cron backfill mechanisms for missed events.
   - Optimized order-placement APIs using Redis Lua script to reduce latency.
   - Stack: Ruby, Rails, Go, Kubernetes, AWS, PostgreSQL, Redis.

2. KreditBee | Software Engineer (Full-time) | June 2023 - June 2025 | Onsite
   - Built underwriting rule engine in microservices architecture.
   - Developed 5+ REST APIs with vendor integrations (TransUnion, PayU, MobiKwik).
   - Reduced response time from 5s to 3s and improved throughput using AsyncIO.
   - Increased loan disbursal from 60 Cr to 120 Cr/month and reduced approval time by 17%.
   - Developed OCR service with Google Cloud Vision to automate document verification.
   - Reduced manual intervention by 15%, saving around 12 lakhs annually.
   - Stack: Python, AWS, Django, FastAPI, Redis, Git, MySQL.

3. KreditBee | Software Developer Intern | Feb 2023 - June 2023 | Onsite
   - Built backend workflow for bank account approval in loan disbursals.
   - Added fraud checks using tampering detection, AML signals, and blacklist data.
   - Automated 350+ account approvals/day and identified 30+ blacklist profiles/day.

Projects
1. Reviews AZ (NLP, Web Scraping)
   - Sentiment analysis over 10,000+ Flipkart reviews.
   - Achieved 93% prediction accuracy with classification models.
   - Tools: Python, Django, BeautifulSoup, Bag of Words.

2. Human-Machine Interaction Anti-Spoofing Face Recognition
   - Built deep learning-based anti-spoofing checks for 2D image/video replay attacks.
   - Used MediaPipe facial landmarks and random instruction strategy.
   - Tools: Python, TensorFlow, MediaPipe, MTCNN.

3. WhichBook (Recommendation System)
   - Implemented popularity and collaborative filtering model using Kaggle data.
   - Built Flask REST API and frontend for user recommendations.
   - Tools: Python, Flask, HTML, CSS.

Honors
- Best New-comer Award, KreditBee - Feb 2023
- Star Performer Award, KreditBee - Sept 2024
- GATE 2021 All India Rank 137