Headline

Software Development Engineer | Golang, Python, AWS, Kafka, Flutter | HLS Streaming, Microservices, AI, MCP, Agent

About

I’m a Backend Software Engineer with 2+ years of experience building and scaling low-latency, high-throughput backend systems using Golang, AWS, and distributed microservices. I’ve worked on performance-critical products where reliability and scale matter, including reducing API p95 latency from 7 seconds to 10 ms, re-architecting video streaming pipelines to cut processing failures by 70%, and designing real-time leaderboard and reward systems supporting 100K+ concurrent users. My experience spans Redis, Kafka, FFmpeg, CloudFront, IVS, PostgreSQL, and event-driven architectures, with a strong focus on observability, cost optimization, and system resilience. Previously, I built high-frequency Golang microservices for real-time location broadcasting, implemented Kafka-based speech-to-text pipelines backed by custom ML models, and deployed computer-vision services using FastAPI and PyTorch. I’ve also delivered multiple Flutter applications used by 10K+ daily users, giving me a strong product mindset and end-to-end ownership perspective. I enjoy solving complex backend problems, optimizing systems at scale, and building clean, maintainable architectures, and I’m currently open to backend or platform engineering roles where I can work on challenging distributed systems and performance-driven products

Location

Bengaluru

AI stack

antigravity

claude code

claude

cursor

langchain

claudecode

Work experience

Software Engineer 2 at Loco, Mar 2025 - Present · 1 y 6 mo, Bengaluru, Karnataka, India · Hybrid

Software Development Engineer (Backend) at Loco, Mar 2025 - Jan 2026 · 11 mos, Bengaluru, Karnataka, India · Hybrid, • Feed API p95 latency from 7 sec to 10 ms by diagnosing Redis sorted-set bottlenecks and optimizing it • Reduced log storage by 50% through log deduplication,index reduction, improved observability practices • Contributed to the monolith-to-microservices migration by developing the feeds Content Service • Re-architected the HLS - MP4 video pipeline, reducing processing failures by 70%, enforcing FFmpeg CPU limits, eliminating pod crashes, and avoiding a costly aws MediaConvert migration • Built a scalable Leaderboard service supporting 100K+ concurrent users and 1K active leaderboards, handling scheduling, scoring aggregation, winner computation, rewards allocation, and sending notifications • Implemented a real-time reward assignment system using an in-house notification service, scaling to 50K+ concurrent users and integrated with the platform’s reward engine. • Secured CloudFront + IVS playback API, preventing misuse and saving $10K/day in IVS costs

Software Developer at 6th Labs, Apr 2023 - Feb 2025 · 1 y 10 mo, India, Architected and deployed high-performance microservices backend in Golang, handling 1000+ concurrent user locations with gRPC streaming and Kalman filtering, achieving 99.9% uptime Developed and published cross-platform Flutter applications with background location tracking and gRPC integration, reducing battery consumption by 20% while maintaining seamless offline synchronization Implemented real-time data processing pipeline using Kafka and Python consumers, integrating Wav2Vec speech-to-text models for audio stream processing at scale Led AWS infrastructure deployment with EC2, load balancing, and auto-scaling groups, ensuring robust scalability and establishing comprehensive CI/CD pipelines that improved development efficiency by 30% Spearheaded development of AI-powered traffic violation detection system using YOLO models, processing thousands of images daily with 95% accuracy and automated reporting Integrated ChromaDB for vector-based search capabilities, enabling AI-driven recommendation systems that increased product discovery rates by 35% Architected and deployed WhatsApp/Telegram sales bot integrating LLM models and RAG for intelligent product discovery, resulting in 40% increase in automated customer interactions Optimized database performance through advanced indexing and query optimization, reducing API response times by 40% while maintaining data integrity across services

Research Intern at Ahmedabad University, Nov 2020 - Feb 2021 · 4 mos, Ahmedabad, Gujarat, India, During my tenure at Ahmedabad University, I undertook a research internship that honed my skills in data processing and computer vision. My responsibilities included: Data Acquisition and Labeling: I developed web scraping tools to collect extensive datasets, which I meticulously labeled to ensure high-quality inputs for model training. Model Fine-Tuning: Leveraging PyTorch, I fine-tuned the YOLOv5 model for object detection tasks, adapting it to our specific dataset to enhance its performance. Model Evaluation: I conducted comparative analyses of various text recognition models, assessing their accuracy and efficiency to identify the most suitable model for our applications. Object Classification: I implemented and evaluated different object classification algorithms to improve the system's ability to accurately categorize detected objects. This experience provided me with a solid foundation in computer vision techniques and practical expertise in utilizing PyTorch for model development and evaluation.

Education

Bachelor of Engineering - BE, Computer Science, Ahmedabad University

Public links

github: https://github.com/dhruv-kabariya