All mentors

Suiyao Chen

Applied Scientist at Amazon | Ph.D. | ML, LLM, GenAI, Agentic, Search, Recsys, Data Science

Mountain View, United States (-08:00 UTC) English, Chinesefrom Mountain View, United States
Usually responds in 1 day
Free
Price per hour
15 min
Time Blocks Available
5.00
5 reviews / 9 sessions
Fri
16
Next availability

Bio

Breaking into top-tier Applied Scientist and MLE roles requires more than just coding skills—it requires deep statistical intuition and the engineering maturity to take models to production. I am an Applied Scientist with 6+ years of industrial experience and a 14-year background in statistical modeling and advanced analytics. I specialize in helping professionals bridge the gap between 'running experiments' and building scalable, production-grade AI systems (including LLMs and Generative AI). I primarily mentor aspiring Applied Scientists, Data Scientists and MLEs who want to: Crack the Interview: Master the specific, high-bar loops for Applied Scientist and MLE roles—focusing on ML System Design, GenAI/LLM concepts, and statistical fundamentals. Transition to Applied Science and Engineering: Move from pure analytics/research into building end-to-end production pipelines. Deepen Technical Skills: Go beyond the basics of model training into advanced fine-tuning, system architecture, and real-world deployment strategies. I created TalentMe (https://talentme.airsota.com/), offering coaching with an AI-native experience that integrates all rich and comprehensive knowledge structures and domain expert insights into your personal AI agent. TalentMe identifies your gaps, generates focused preparation briefs, and transforms every mock interview into measurable progress.

Expertise

  • Artificial intelligence

    I am an Applied Scientist with 6+ years of industrial experience and a 14-year background in statistical modeling and advanced analytics. I specialize in helping professionals bridge the gap between 'running experiments' and building scalable, production-grade AI systems (including LLMs and Generative AI).

  • Data science

    I am an Applied Scientist with 6+ years of industrial experience and a 14-year background in statistical modeling and advanced analytics. I specialize in helping professionals bridge the gap between 'running experiments' and building scalable, production-grade AI systems (including LLMs and Generative AI).

Experience

  • Amazon

    Applied Scientist, Core Search

    ◦ Lead a team of scientists and MLEs to built RAG-based brand detection system for query understanding with LLM-finetuning and achieve 400M+ annualized revenue gain and reduce over 80% operation cost with pipeline simplication. ◦ Architected and delivered a novel Agentic Search \& Personalization Framework pioneering 4B-parameter models fine-tuning through SFT and DPO to outperform Claude Sonnet 4.5 and defined the team's evaluation roadmap and set new standards for productionizing high-performance small language models. ◦ Spearheaded the technical architecture of a Multi-Modal LLM Indexing Service, driving a paradigm shift in search product discovery and scaling to process millions of products with \$600M+ annual revenue gain. ◦ Engineered a Multi-Modal Retrieval System by designing an LLM-driven synthetic data pipeline to train semantic matching models, enabling robust hybrid image-text search capabilities.

  • Amazon

    Applied Scientist, Fraud/Abuse Detection

    ◦ Designed and pre-trained a 1B-parameter Transformer-based foundation model on multi-modal customer behavior sequences, generating robust embeddings that power downstream fraud detection tasks. ◦ Built abuse risk detection models and improve AUC from 75% to 98% and recall from 50% to 80%. ◦ Developed deep tabular representation learning frameworks to facilitate automatic feature engineering and improve downstream classification models by 5% in $Recall and 10M+ savings.

  • Amazon

    Data Scientist, Fraud/Abuse Detection

    ◦ Designed a graph-based data pipeline to evaluate customer risk propensity through linkage message passing and deployed with 1000+ variables for worldwide coverage and 90% cost reduction after migration. ◦ Designed an automated self-service variable monitoring system for in-time monitoring with scalability and flexibility to reduce the monitor onboarding efforts from 2 months to 1 day.

  • IBM

    Senior Data Scientist
    •ibm.com

    Team: Offering and Marketing Analytics Leveraged 200K+ reviews for 1K+ products to support the strategic decision making for offering managers, sellers and marketing teams. Developed model pipeline with strategic web-scraping, sentiment analysis, topic classification using NLP and provide actionable insights. Developed graph analytics framework for product competition network analysis with exceptional visualization and explainability Team: Sales Channels Effectiveness Developed a deep dive solution on business transformation and collaborated with cross-functional teams in Finance and Global Sales. Consolidated data from various sources and performed data analysis for revenue trends and ranking under business constraints. Delivered strategies with integrated client data and structured tiering for global client segmentation and sales resources allocation.

  • University of South Florida

    Teaching Assistant

    •Spring20 ESI6340 Probabilistic Operations Research •Fall19 ESI4606 Engineering Analytics •Spring19 EGN3443 Probability and Statistics for Engineers •Fall18 EGN3443 Probability and Statistics for Engineers •Spring18 EGN3443 Probability and Statistics for Engineers •Fall17 ESI6247 Statistical Design Models •Spring17 EIN6936 Engineering Analytics II •Fall16 ESI4606 Engineering Analytics I •Summer16 EGN3615 Engineering Economics •Spring16 EGN3615 Engineering Economics

  • University of South Florida

    Research Assistant

    Research Interests & Experiences: applied statistics, Bayesian data fusion, machine learning, big data, survival analysis, reliability and quality assurance, risk assessment, healthcare.

  • IBM

    Senior Data Scientist Intern - Chief Analytics Office
    •ibm.com

    Built cognitive pricing models for business consulting project and automate manual negotiation process. Led the duties of data consolidation, feature engineering, predictive modeling and price optimization, with personalized pricing guardrails and innovation on strategic values for gross profit maximization. Contributed to developing real-time pricing API in Python and interactive dashboards for stakeholders. Delivered strategic decision making plans with over $80 million expected impacts on global business.

  • Express Scripts

    Data Science Intern

    Built comprehensive framework of predictive modeling and decision making for pharmacy claims audit optimization project with over $22 million expected benefits in business development for the company. Applied machine learning techniques to high dimensional data with 850 columns and achieved over 84% prediction accuracy. Facilitated daily-based predictive scoring and efficient discrepancy auditing for millions of prescriptions from thousands of pharmacies in the network.

  • Statistical and Applied Mathematical Sciences Institute

    Industrial Math/Stat Modeling (IMSM) Workshop

    Group project: ``Big Data and Data Fusion: Exploring EPA’s Air Quality Products and Remote Sensing Options with Cloud Computing". Conducted group project with big data (sensor data, satellite images, etc.) to explore and evaluate remote sensing products from U.S. Environmental Protection Agency, finished with presentation and final report. Performed data fusion, geographical visualization using Google Earth Engine with JavaScript and data exploration, clustering and time series analysis through R and Python.

  • Columbia University in the City of New York

    Office Assistant

    Executed duties of reception, data entry, maintaining event files, bookkeeping and sign making. Coordinated with office associates to deal with daily reservations and complex teamworks with details.

  • Centaline Property

    Marketing and Strategic Research Assistant

    Determined survey strategy on real estate market by targeted interviews and random interviews. Analyzed real estate data using regression models and market situations using Porter Five Forces Models. Reported the research result of the property market and explored new business plans based on data analysis.

Made within Glyfada