Aniket Sood

About Aniket Sood

Passionate software engineer with a strong foundation in backend development, scalable system design, and cloud-native architectures. Currently working as a Senior Software Engineer at Samsung Research Institute, where I architect data-driven microservices powering ML platforms used by 300+ engineers.> Recieved Q1, 2025,\"Samsung SPOT Award\" for - Gallery Query Processor SLM Encoder data enhancements, end to end training, evaluation and deployment.> Recieved Q2, 2024,\"Special Recogntion\" Award for - Building backend databse collection part for chat-gpt like LLM, trained on specific 5gFine-tuned lightweight language models (MiniLM v21) using LoRA for semantic query understanding in gallery search, improving intent extraction accuracy by 15%.(Java, Python, HuggingFace, LoRA, Transformers). Integrated trained models into Android apps, performing code-level changes, debugging crashes, and improving inference stability across 3+ internal clients. Previously at Rymo, I built an Android app for pose classification using MediaPipe BlazePose, optimized Bluetooth comms via BLE, and significantly reduced latency. Key Strengths:System & Low-Level DesignEvent-Driven MicroservicesScalable API Design (Java, Spring Boot)AWS Ecosystem (EC2, API Gateway, SQS, ElastiCache)Full-Stack Tools: React, Node.js, Redis, MongoDBML Integrations & Real-Time Analytics Tools Projects like Stock-News Tracker, TinyURL at scale, and School Dashboard demonstrate my ability to engineer complete, production-level systems. Let’s connect if you’re building complex distributed systems, enabling AI at scale, or just love talking clean architecture and debugging weird edge cases.

Experience

1. Senior Software Engineer

   Samsung Electronics

   Mar 2025 — Present · Bengaluru, IN

   Part of Platform Intelligence team, Al productization part> Recieved Q1, 2025,\"Samsung SPOT Award\" for - Gallery Query Processor SLM Encoder data enhancements, end to end training, evaluation and deployment.Architected and implemented Telemetry Data Aggregator service, processing 50M+ events/day, achieving 99.98% uptime and reducing data ingestion latency by 43%.Fine-tuned lightweight language models (MiniLM v21) using LoRA for semantic query understanding in gallery search, improving intent extraction accuracy by 15%.(Java, Python, HuggingFace, LoRA, Transformers). Integrated trained models into Android apps, performing code-level changes, debugging crashes, and improving inference stability across 3+ internal clients- Working on various NLP classification related tasks- Working with transformers for various data generation and classification tasks- Working on finetuning and transfer learning of the state of the art transformer models, like MT5, and with fine tune algo LoRA

Education

• Punjab Engineering College

  Bachelor of Technology - BTech, Electronics and Communications Engineering

  2019 — 2023

• Sri Guru Harkrishan Model Public School, Sec-38 D,Chd

  Both Sec and Sr.Sec

  2015 — 2018