Siffi Singh

I am an Applied Scientist at Amazon Web Services (AWS) in Los Angeles, USA. I work with Amazon Connect and Amazon Transcribe teams building multilingual conversational and generative AI systems for speech understanding and analytics. Prior to this, I spent several years on the Amazon Alexa team in Camridge, UK working on automated knowledge graph construction from web-scale information.

Beyond industry work, I collaborate with researchers exploring the intersection of fundamental science and artificial intelligence. I work on research projects with Prof. Sergei Gukov at the California Institute of Technology and previously with Prof. Burigede Liu at the University of Cambridge.

I received my Masters of Science from National Chiao Tung University (NCTU) under the supervision of Prof. Hsueh-Ming Hang. I previously earned a Bachelor of Technology with distinction in Computer Science from Dr. APJ Abdul Kalam Technical University. Following my undergraduate studies, I worked as a Researcher with Prof. Carrson C. Fung at NCTU, Taiwan and later interned at the Industrial Technology Research Institute (ITRI) developing computer vision for autonomous driving vehicles.

Email CV Scholar LinkedIn GitHub

Siffi Singh
AWS Alexa Caltech Cambridge NCTU AKTU ITRI

News §

Research §

Automatic Speech Translation Automatic Speech Translation
Amazon Transcribe, AWS
We enable real-time speech translation across 10+ languages, powering multilingual call centers, healthcare consultations, and emergency services.
Generative Trend Detection Generative Trend Detection
Amazon Connect Contact Lens, AWS
We enable continuous detection of emerging customer issues using generative AI, tracking trends across conversations over time to power smarter contact center insights for global enterprises.
Multilingual PII Redaction Multilingual PII Redaction
Amazon Connect Contact Lens, AWS
We enable robust multilingual PII redaction using AI, protecting sensitive data across global call and chat interactions for enterprise customers.
AWS Launch
Multilingual PII Redaction Conversational Theme Detection
Amazon Connect Contact Lens, AWS
We enable automated theme detection across customer conversations using AI, uncovering key drivers of contact at scale for enterprise contact centers.
AWS Launch | Docs
Conversational Semantic Matching Conversational Semantic Matching
Amazon Connect Contact Lens, AWS
We deliver intent-aware semantic search across customer conversations, going beyond keywords to surface relevant insights with high precision across call and chat data.
AWS Launch
Structured Knowledge Extraction Structured Knowledge Extraction
Alexa Information Domains
We extract structured knowledge from web tables using semi-supervised learning, scaling knowledge graph construction by aligning tabular data to real-world entities and relations.
Relation Extraction Paper | Column Class Detection Paper
RedTab Tabular Benchmark Dataset RedTab Tabular Benchmark Dataset
Alexa Information Domains
We built a large-scale annotated benchmark for tabular understanding, enabling robust evaluation and advancement of column classification and relation extraction models.
RedTab Paper
Knowledge Graph Expansion Knowledge Graph Expansion
Alexa Information Domains
We built scalable pipelines to ingest and refresh structured knowledge, expanding Alexa's knowledge graph with millions of high-quality facts while maintaining accuracy over time. Launch

Invited Talks §

REWORK London AI Summit 2022 “Making Alexa More Knowledgeable”
REWORK London AI Summit (November 2022)
Event Page
AI Expo London 2022 “Making Alexa More Knowledgeable”
AI & Big Data Expo Global, Olympia London (June 2022)
Speaker Profile  |  LinkedIn Post

Podcasts §

I run Minds in Motion, a platform dedicated to conversations with researchers, scientists, and thinkers about what truly drives them. It is a space for honest perspectives on research, career transitions, new beginnings, and the question that connects them all "what keeps you going?"

YouTube  |  LinkedIn  |  Instagram

Back to top