Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I’m Avi Kumar Talaviya, a Machine Learning Engineer and Java Full Stack Developer based in San Francisco, open to relocation to Seattle, Plano, Chicago, or New York. I specialize in building AI-powered backend services, distributed systems, and high-scale platforms that drive tangible business impact. With 3+ years of experience at Meta and Accenture, I excel in Java 21/17, Python, Spring Boot, Kafka, and React/Next.js to deliver low-latency, highly available systems serving 15M+ devices, 500M+ daily events, and 50M+ transactions/day. I love collaborating with AI researchers, product, and operations teams to advance AI personalization, AdTech monetization, fraud detection, and cloud-native architectures on AWS, Azure, and Kubernetes.…Hi, I’m Avi Kumar Talaviya, a Machine Learning Engineer and Java Full Stack Developer based in San Francisco, open to relocation to Seattle, Plano, Chicago, or New York. I specialize in building AI-powered backend services, distributed systems, and high-scale platforms that drive tangible business impact. With 3+ years of experience at Meta and Accenture, I excel in Java 21/17, Python, Spring Boot, Kafka, and React/Next.js to deliver low-latency, highly available systems serving 15M+ devices, 500M+ daily events, and 50M+ transactions/day. I love collaborating with AI researchers, product, and operations teams to advance AI personalization, AdTech monetization, fraud detection, and cloud-native architectures on AWS, Azure, and Kubernetes.

Avi Kumar Talaviya

AI Engineer, Full Stack Developer, Back-End Developer, +2





Hi, I’m Avi Kumar Talaviya, a Machine Learning Engineer and Java Full Stack Developer based in San Francisco, open to relocation to Seattle, Plano, Chicago, or New York. I specialize in building AI-powered backend services, distributed systems, and high-scale platforms that drive tangible business impact. With 3+ years of experience at Meta and Accenture, I excel in Java 21/17, Python, Spring Boot, Kafka, and React/Next.js to deliver low-latency, highly available systems serving 15M+ devices, 500M+ daily events, and 50M+ transactions/day. I love collaborating with AI researchers, product, and operations teams to advance AI personalization, AdTech monetization, fraud detection, and cloud-native architectures on AWS, Azure, and Kubernetes.…Hi, I’m Avi Kumar Talaviya, a Machine Learning Engineer and Java Full Stack Developer based in San Francisco, open to relocation to Seattle, Plano, Chicago, or New York. I specialize in building AI-powered backend services, distributed systems, and high-scale platforms that drive tangible business impact. With 3+ years of experience at Meta and Accenture, I excel in Java 21/17, Python, Spring Boot, Kafka, and React/Next.js to deliver low-latency, highly available systems serving 15M+ devices, 500M+ daily events, and 50M+ transactions/day. I love collaborating with AI researchers, product, and operations teams to advance AI personalization, AdTech monetization, fraud detection, and cloud-native architectures on AWS, Azure, and Kubernetes.

Available to hire

Hi, I’m Avi Kumar Talaviya, a Machine Learning Engineer and Java Full Stack Developer based in San Francisco, open to relocation to Seattle, Plano, Chicago, or New York. I specialize in building AI-powered backend services, distributed systems, and high-scale platforms that drive tangible business impact.

With 3+ years of experience at Meta and Accenture, I excel in Java 21/17, Python, Spring Boot, Kafka, and React/Next.js to deliver low-latency, highly available systems serving 15M+ devices, 500M+ daily events, and 50M+ transactions/day. I love collaborating with AI researchers, product, and operations teams to advance AI personalization, AdTech monetization, fraud detection, and cloud-native architectures on AWS, Azure, and Kubernetes.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Work Experience

AI/ML Engineer at Meta

August 1, 2025 - Present

Architected AI-powered backend services for Meta Smart Glasses supporting 15M+ devices, reducing p99 latency by 38% and achieving 99.99% service availability. Optimized AI, media, and AdTech systems to 2.7× throughput, cut infrastructure costs by 22%, and improved sponsored recommendation CTR by 7.5%. Collaborated with AI researchers and product teams to deliver multimodal AI experiences with voice, vision, and personalization, processing 500M+ daily events. Developed Java 21, Spring Boot, and Spring Cloud microservices with REST/gRPC, Redis, and PostgreSQL for device management and personalization, handling 1M+ requests/hour. Engineered Kafka-based media pipelines for photo uploads, video streaming, and metadata processing, handling 250M+ events/day with 99.8% reliability. Integrated Meta AI, Llama, STT, and Vision AI services, reducing AI response latency by 35% and improving recommendation accuracy by 12%. Built AdTech services for sponsored recommendations, user segmentation, and

Software Engineer at Accenture India

March 1, 2023 - June 1, 2025

Developed Java 17/Spring Boot microservices for payment authorization, settlement, and billing systems processing 50M+ transactions/day with 99.99% SLA. Designed Kafka-based event-driven architectures for payment, refund, and reconciliation workflows, improving throughput by 2.3× and reducing failure recovery time by 40%. Built secure REST APIs using Spring Cloud, Redis, and PostgreSQL, reducing p99 latency by 35% and improving transaction success rates by 15%. Implemented Saga patterns, idempotency, and distributed transactions to ensure reliable payment execution and exactly-once processing across microservices. Developed fraud detection and risk-scoring services using Kafka Streams, Redis, and Cassandra for sub-500ms analysis, enabling real-time monitoring and reduced investigation time by 30%. Engineered revenue reconciliation and settlement pipelines using Spring Batch, Oracle, and PostgreSQL, reducing reconciliation turnaround by 20%. Built React and TypeScript dashboards for tr