Senior AI/ML Consultant · AWS Professional Services

Ardya Dipta Nandaviri

AI leader specializing in generative and agentic AI, with 10+ years of experience designing and deploying scalable, production-grade intelligent systems—from LLM-powered agents and RAG applications to reinforcement learning, recommendation platforms serving 10M+ daily users, and large-scale fraud detection. Carnegie Mellon Robotics alum and Senior AI/ML Consultant at AWS Professional Services, based in Singapore.

Credentials

Certifications and recognition. AWS badges are verifiable on Credly.

AWS Certified Machine Learning – Specialty AWS Certified Machine Learning Engineer AWS Certified AI Practitioner AWS Certified Cloud Practitioner Google Certified GenAI Leader Google Certified TensorFlow Developer Google Developer Expert, AI/ML & GenAI (2024–2025) Harvard Business School Online, Negotiation Mastery (2023)

Experience

From robotics perception at Carnegie Mellon and Caterpillar to leading data science organizations at Indonesia's largest tech and healthcare companies.

Senior AI/ML Consultant · Amazon Web Services

Oct 2025 – present

ASEAN Professional Services, Singapore

  • Helping customers across Southeast Asia accelerate business growth through AI/ML, from generative AI and agentic applications to production ML platforms.

Head of Data Science and Business Insights · Kalbe Farma

Jun 2023 – Jul 2025

largest pharmaceutical, healthcare and nutrition company in SEA, ~$4B valuation

  • Spearheaded data science strategy and end-to-end ML deployment, leading the team to deliver 20B IDR in revenue growth with reinforcement learning for personalized marketing and recommendations.
  • Cut overstock by 30% with time-series forecasting models optimized for inventory management.
  • Built and deployed an LLM chatbot for business inquiries using RAG and text-to-SQL.
  • Led bioinformatics data scientists to research and deploy a genomics-based polygenic risk score (PRS) algorithm, improving the accuracy of Kalbe's Nutrigenme genomic testing.

Lead Data Scientist, Fraud & Recommendation Systems · Gojek

Jan 2021 – Jun 2023

Indonesia's first tech unicorn, ride-hailing super-app, ~$5.1B valuation

  • IC and team lead; mentored 3 data analysts and 4 data scientists, 4 of whom were promoted.
  • Recommendation system: ~$1M monthly revenue growth from an end-to-end MLOps pipeline serving 10M+ DAU, lifting CTR 10% with multi-armed-bandit personalization and cutting deployment latency 40%.
  • Fraud detection: potentially saved $3M+ in six months with a TensorFlow autoencoder anomaly detector that cut incentive abuse by 80%+; reduced scams from illegal and fake-GPS apps by 90%+ by classifying GPS ping patterns.

Senior Data Scientist · Oct 2017 – Dec 2020

  • Architected the fraud detection system (TensorFlow autoencoder) that removed 90%+ of incentive abusers.
  • Designed scalable MLOps for routing optimization and churn prediction at 10M daily active users.
  • Integrated Google Cloud Vertex AI for real-time predictions with a <200 ms SLA.

Lead Data Scientist · Amartha

Nov 2016 – Oct 2017

peer-to-peer lending, $53M Series C

  • Led the data science team and deployed a credit scoring system that held non-performing loans at ≤ 0.5%.

Robotics Research Programmer & Analyst · Carnegie Mellon University, The Robotics Institute

Jan – Oct 2016

Pittsburgh

  • Developed and tested software for self-driving vehicle projects, including an Extended Kalman Filter sensor-fusion system. C++, MATLAB, ROS.

Robotics Perception Engineer · Caterpillar Inc.

Jan – Dec 2015

Pittsburgh

  • Built ML and computer vision systems to improve obstacle detection for an autonomous mining truck. C++, Python, Linux.
The World BankData Scientist (part-time) · 2018–2019
Lifeshel LLCSoftware & Electrical Engineer · 2014, USA
IBMIT Specialist · 2010–2013, Indonesia

Projects

Open-source work and write-ups. Private repositories are available on request for hiring purposes.

LLM chatbot assistant

Document-querying chatbot built with Google Gemini, FAISS and LangChain to improve employee productivity.

GeminiFAISSLangChain

RLHF for GPT

Fine-tuning GPT models with reinforcement learning from human feedback to improve user engagement metrics.

RLHFFine-tuningPyTorch

Lie detection from micro-expressions

Computer vision and deep learning system that detects deception from facial micro-expressions. Top-10 finalist, Bank Indonesia AI Hackathon 2024.

Computer visionTensorFlowVertex AI

Fraud detection in health insurance

ML pipelines in Python, PyTorch and GCP that reduced fraudulent claims by 15% for one of Indonesia's largest insurers.

PyTorchGCPImbalanced learning

Perception for self-driving cars

Perception and sensor-fusion work from Carnegie Mellon's Robotics Institute, with an accompanying paper.

RoboticsSensor fusionC++

Google Lunar X-Prize moon lander

Robotics work on a lunar lander robot for the Google Lunar X-Prize, 2013–2014.

Space roboticsCMU

Writing

Field notes on production AI, retrieval, and machine learning systems.

Latest · Long-form guide with interactive walkthrough

From 100,000 documents to grounded answers

A long-form design walkthrough: tiered ingestion (layout parsing, OCR, vision captions), structure-aware chunking, metadata filtering, BM25 and HNSW retrieval, reciprocal rank fusion with reranking, and when GraphRAG earns its cost. Paired with an interactive, animated 12-step explainer.

RAGHybrid retrievalGraphRAGSystem design

Achievements

Competitions, awards and recognition.

  • 2010Ganesha Karya Award from the Rector of ITB, top 10 of 3,000 students
  • 2009Champion and Best Algorithm, Indonesia National Intelligent Robot Contest, beating 30+ universities (Kompas, national TV)
  • 2009National Champion, Smart Robotic Contest, Ministry of Education of Indonesia
  • 2020–Mentor, Bangkit ML bootcamp by Google, Gojek, Tokopedia and Traveloka

Skills

ML techniques

Reinforcement learningRecommendation systemsLLMs & GenAIRAG & agentsComputer visionMLOpsFraud detectionImbalanced learningTime-series forecastingConvex & non-convex optimization

Languages

PythonC++SQLJavaR

Frameworks & platforms

TensorFlowPyTorchDockerKubernetesAirflowSparkAWSGoogle Cloud PlatformVertex AI

Education

Carnegie Mellon University

M.S. in Robotic Systems Development, GPA 3.50 · Pittsburgh, 2013–2014

Machine Learning 10-601 (Tom Mitchell) · Computer Vision 16-720 (Martial Hebert)

Institut Teknologi Bandung

B.Sc. in Electrical Engineering · Bandung, 2006–2010

National Champion, Smart Robotic Contest (Ministry of Education, 2009)

Harvard Business School Online

Certificate in Negotiation Mastery · 2023

Contact

Open to conversations about applied AI, ML platforms, and speaking engagements.