Open to opportunities

Divya Sangwan

@divyasangwan

Message

Software engineer and ML builder focused on scalable systems and measurable model impact.

India

Message

What I'm looking for

I’m looking for a role where I can build scalable backend and ML systems end-to-end—owning design, shipping production improvements, and collaborating with a team that values reliability, modular architecture, and measurable impact.

I’m a Software Engineer building production-grade, scalable backend services and ML-driven solutions. At Arcesium, I won 1st place in a 30-hour hackathon by developing an ML-based fuzzy matching system and then turned that engineering mindset into reliable distributed features.

I’ve worked across distributed systems and applied machine learning—implementing distributed locking with Zookeeper (Kotlin), designing in-memory caching with Hazelcast to cut API latency by ~85%, and creating modular backend architectures from scratch. I also strengthened system stability by troubleshooting real production issues and improving throughput for data-intensive workflows.

From Amazon as an Applied Scientist Intern to building low-latency APIs and real-time streaming pipelines with Flask and Apache Kafka, I enjoy owning the full stack of problem-solving. My research project with CSIRO focused on anti-phishing defenses using visual similarity models (pHash/dHash and Siamese networks), and I’ve also shipped applied computer-vision work like traffic congestion control using yolov8 and Deepsort.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer

Current

Arcesium

Jul 2024 - Present (1 year 11 months)

Won 1st place in a 30-hour hackathon by building an ML-based fuzzy matching system. Implemented Zookeeper distributed locking and Hazelcast caching, reducing API latency by ~85% and improving replica synchronization reliability in concurrent workloads.

Kotlin Apache Zookeeper Hazelcast Distributed Locking Backend Services Distributed Systems Production Troubleshooting

Applied Scientist Intern

Amazon

Jan 2024 - Jun 2024 (5 months)

Developed features to mitigate marketplace abuse by integrating XGBoost ensembles and GNN models, improving precision by 9% (XGBoost) and 8% (GNN) over baselines. Used LLM-driven synthetic data generation for imbalanced datasets and deployed GNN models via SageMaker for the US marketplace.

XGBoost LLM Synthetic Data Generation Imbalanced Learning Amazon SageMaker Machine Learning

Full Stack Intern

JM Financials

Jun 2023 - Jul 2023 (1 month)

Built a Python TCP/IP client to receive and process exchange feed data, using multithreading and locks/mutex to achieve 100% data integrity while cutting processing time by 40%. Developed real-time trade data APIs with Flask and integrated Apache Kafka for distributed streaming and synchronization.

Python TCP IP Multithreading Flask REST APIs Apache Kafka Real time Data Processing

Software Developer Intern

Fusion Tech Process Consulting

Jun 2022 - Jun 2023 (1 year)

Built a scalable mail engine microservice using a Pub/Sub architecture for distributed messaging and integrated Node services with Amazon S3, GCP Storage, and threat intel via REST APIs. Strengthened security with firewall rules, static code analysis, and SIEM rule integration for log management, and used PySpark to run SQL in a multi-threaded environment for high scalability.