CV — Sina Torfi

Summary

Senior Applied Scientist with 8+ years of experience building production AI systems at Amazon, Meta, and high-growth startups. Currently contributing to Amazon Q's LLM capabilities — function calling, alignment, and post-training optimization. Research spans large language models, NLP, computer vision, and privacy-preserving AI, with 15+ peer-reviewed publications (1,000+ citations) and 40,000+ GitHub stars across open-source projects. Ph.D. in Computer Science from Virginia Tech.

Professional Experience

Senior Applied Scientist

Amazon — Amazon Q

Jan 2025 – Present
San Jose, CA

Design and implement advanced function-calling and tool-use capabilities for Amazon Q, enabling multi-step agentic reasoning and orchestration across enterprise knowledge sources.
Develop post-training pipelines including instruction tuning, DPO-based preference optimization, and targeted fine-tuning strategies that improve response quality, factual grounding, and task completion in complex enterprise workflows.
Research and deploy privacy-preserving inference techniques and alignment methods — including constitutional AI guardrails, PII-aware generation, and safety classifiers — for trusted enterprise LLM deployment at scale.
Build comprehensive evaluation frameworks for agentic LLM behaviors, incorporating multi-turn reasoning benchmarks, function-calling accuracy metrics, and automated red-teaming to drive rapid iteration cycles.

Research Scientist

Meta — Reality Labs & Recommendations

May 2022 – Dec 2024
Redmond, WA

Designed and shipped multimodal foundation model architectures for AR/VR, achieving 0.96 mIoU on semantic segmentation through novel SWIN-based Vision Transformer pipelines with knowledge distillation and quantization-aware training.
Built cross-modal knowledge transfer systems using CLIP and cross-attention mechanisms for VQA and image captioning, enabling zero-shot generalization across unseen domains.
Developed large-scale deep sequence recommendation models serving billions of daily predictions, incorporating Transformer-XL architectures with meta-learning for cold-start optimization.
Drove research on self-supervised pre-training, contrastive learning, and domain-adaptive strategies adopted across multiple product surfaces.

Senior Data Scientist

Huma AI

Jun 2020 – May 2022

Architected end-to-end cancer detection pipelines combining hierarchical CNNs, attention mechanisms, and self-supervised pre-training on pathology images, deployed on AWS SageMaker for clinical-scale inference.
Invented differentially private synthetic data generation frameworks using convolutional GANs and diffusion models, enabling HIPAA-compliant model training (published in Information Sciences, 140+ citations).
Designed adversarial debiasing and causal inference frameworks (Double ML) for disease prediction, reducing diagnostic bias while maintaining clinical accuracy.

Graduate Research Assistant

Virginia Tech / West Virginia University

Jan 2016 – Aug 2020
Blacksburg, VA

Published 15+ peer-reviewed papers in top venues (IEEE Access, Information Sciences, arXiv) spanning NLP, computer vision, generative models, and audio-visual learning — accumulating 1,000+ citations.
Pioneered 3D CNN architectures for audio-visual speaker recognition and lip reading, establishing new benchmarks (150+ citations, IEEE Access 2017).
Authored a comprehensive NLP deep learning survey that became a widely referenced resource in the field (380+ citations).
Created high-impact open-source projects (TensorFlow-World, lip-reading-deeplearning, speechpy) with 8,000+ combined GitHub stars.

DSP Engineer

Stealth Company

Sep 2012 – Dec 2014

Designed and optimized real-time digital signal processing algorithms for embedded systems, including adaptive filtering and spectral analysis modules.
Built low-latency audio processing pipelines for production hardware with strict throughput and memory constraints.

Education

Ph.D. in Computer Science

Virginia Tech

2016 – 2020

M.Sc. in Electrical Engineering

West Virginia University

2014 – 2016

Technical Skills

LLM & NLP

Large Language ModelsFunction CallingPost-Training (DPO/KTO/RLHF)AlignmentRAGTransformersInstruction TuningConstitutional AIRed-Teaming

ML & Deep Learning

PyTorchTensorFlowComputer VisionDiffusion ModelsGANsKnowledge DistillationSelf-Supervised LearningContrastive LearningRecommender Systems

Infrastructure

AWS SageMakerDistributed TrainingSpeculative DecodingQuantization (AWQ/GPTQ)ONNXTensorRTKubernetesDocker

Languages & Tools

PythonC++SQLGitLinuxFAISSWeights & BiasesMLflow

Selected Publications

1

Natural Language Processing Advancements By Deep Learning: A Survey. A. Torfi, R.A. Shirvani, Y. Keneshloo, N. Tavaf, E.A. Fox. arXiv, 2020. 380+ citations

2

Differentially Private Synthetic Medical Data Generation using Convolutional GANs. A. Torfi, E.A. Fox, C.K. Reddy. Information Sciences, 2022. 140+ citations

3

3D Convolutional Neural Networks for Cross Audio-Visual Matching Recognition. A. Torfi, S.M. Iranmanesh, N.M. Nasrabadi, J.M. Dawson. IEEE Access, 2017. 150+ citations

4

CorGAN: Correlation-Capturing Convolutional GANs for Generating Synthetic Healthcare Records. A. Torfi, E.A. Fox. FLAIRS, 2020.

5

GASL: Guided Attention for Sparsity Learning in Deep Neural Networks. A. Torfi, R.A. Shirvani, S. Soleymani, N.M. Nasrabadi. arXiv, 2019.

6

Text-Independent Speaker Verification Using 3D Convolutional Neural Networks. A. Torfi, J.M. Dawson, N.M. Nasrabadi. IEEE ICMLA, 2018.

7

SpeechPy — A Library for Speech Processing and Recognition. A. Torfi. Journal of Open Source Software, 3(27), 2018.

Full list: Google Scholar Profile

Open Source & Community Impact

16.3K

TensorFlow-Course

Comprehensive TensorFlow tutorials adopted by thousands of learners globally.

7.1K

machine-learning-course

End-to-end ML course with interactive Python examples.

4.6K

deep-learning-roadmap

Curated roadmap covering DL foundations to cutting-edge research.

4.5K

TensorFlow-World

Research-backed TensorFlow library and educational resources.

40,000+ combined GitHub stars across personal and organizational repositories.