Amin Dorostanian
Senior Machine Learning Engineer
Summary
As a Machine Learning Engineer with 15 years of experience, I specialize in developing efficient solutions, from initial prototyping to production deployment. My approach focuses on creating products driven by clear objectives, skillfully balancing immediate successes with strategic, long-term goals. I am committed to continuous learning, constantly seeking to expand my knowledge. My passion for Artificial General Intelligence (AGI) research and exploring deep philosophical queries is well-known among my close associates.
Machine Learning Engineer | Founder
Neuroship AI (Own Consultancy Company)
2021-02 – Present
Lienden, Gelderland, NETHERLANDS
  • BMW Group: Building GenAI products (Text-to-SQL Assistant), enabling BMW clients to leverage GenAI to query their data more easily. Technologies: RAG, LLMs (open-source and Azure-based OpenAI), AWS ECS, Terraform/CDK, AWS RDS, Snowflake, LLMOPs IaaC, CI/CD pipelines. Gen-AI stack: Python, FastAPI, Fast RAG, Hallucination detection, No Agentic AI Frameworks and pure use of openai lib(they are not helpful)
  • SmartClaim: AI-powered assistant for British R&D tax advisors. Technologies: Full-stack development for seamless UI/UX experience, Svelte, TypeScript, Python, AWS, AWS CDK, Azure, TF-IDF, LSA
  • ECA: Building an Enterprise Coding Assistant—a developer tool that integrates seamlessly with existing workspace apps and leverages Generative AI (LLMs) to enhance software development efficiency. Technologies: Full-stack development (Svelte, Tailwind, Python, Docker, Kubernetes, CI/CD, Terraform, OIDC support, Event-driven architecture)
  • Viasat: Creating templates and solutions for in-house data scientists to leverage the data in the company and create AI-enabled applications. Technologies: Python, AWS SageMaker, AWS CDK, Airflow
  • MN Asset Management: Developing a Data Hub to provide a data-driven architecture enabling various applications. Technologies: Spark, Kafka, Kubernetes, Terraform, Airflow, AWS, Python, Java
  • ABN AMRO Bank: Building a platform for data scientists to bring their models to production. Technologies: Kafka, Azure Cloud Managed Services (Active Directory, etc.), Airflow, Kubernetes, Databricks
  • Deribit: Helping one of the largest derivatives exchanges in the crypto world detect malicious transactions using ML. Analyzing over 100 billion rows using Google BigQuery, ML Engine, and DataFlow on GCP
  • Bitvavo: Developing big data solutions for a Dutch crypto spot exchange to provide customer insights and establish the foundation of their data pipelines. Technologies: Kafka, BigQuery, Python, Kubernetes, CI/CD
  • Neurolytics: MLOps and architectural design for their AI platform
  • Financial Market Analysis: Leveraging near-realtime market data to gain insight and edge over the market. Technologies: Reinforcement Learning (PPO, A3C)
Senior Machine Learning Consultant
Fourkind/ThoughtWorks
2020-09 – 2021-04
Amsterdam, NETHERLANDS
  • Working in various domains to carry out projects using Gaussian Mixture Models and Bayesian Optimization
  • Recipe optimization for Trident gum with a project carried out for Mondelez
  • Involved around all aspects of the life-cycle of the ML projects from modeling to engineering challenges
  • Making business impact by leveraging AI tech
  • Azure Tools, MLOps, Python, Kubernetes
Senior Software Engineer (Operations Decisions Support)
KLM Royal Dutch Airlines
2019-09 – 2020-09
Amsterdam, NETHERLANDS
  • Working on Ground Staff Tasks Optimization Tools
  • Leading software product together with engineering team and Operation Research Consultants
  • Mixed Integer Programming
  • Using Kafka, HDFS, Spark over Python and Scala to build applications for planners to allocate tasks for the workers on the airport
  • Working closely with BCG Consultants on implementing innovative optimizers for Airline Operations
Machine Learning Engineer & Scrum Master
bol.com
2018-07 – 2019-09
Utrecht, NETHERLANDS
  • Working on Intent Recognition for customers questions and conversation analysis
  • Using NLP to comprehend and answer questions in an efficient way to help our customers
  • From prototyping in Python and benchmarking to bringing to production and serving a webservice on Kubernetes
  • Java, Spring Boot, Kotlin, GKE, BigQuery, ElasticSearch, Clustering, TF-IDF, Convolutional Neural Nets
Machine Learning Engineer & Scrum Master
bol.com
2017-12 – 2018-07
Utrecht, NETHERLANDS
  • Working on demand and sales forecasting solutions to make customer experience better for products availability and optimize costs of stocking
  • Big data solutions for building Apache Beam pipelines
Software Engineer
bol.com
2017-05 – 2017-12
Utrecht, NETHERLANDS
  • Working with JVM technologies in a team responsible for shop order intake
  • Based on Spring framework we create our services based on Micro-service design patterns and use the cutting-edge frameworks to deliver high-performance jobs in huge scale
Freelance Data Scientist
Freelance Data Scientist
2016-12 – 2017-06
Stockholm, SWEDEN
  • Working on different projects including retailers data to give insights to the customers
  • Analyzing huge amount of data for online shops to improve recommender systems performance
Machine Learning Engineer
Watty
2016-05 – 2016-12
Stockholm, SWEDEN
  • Working on Deep Learning methods which can be applied on Energy Disaggregation problems
  • Solving challenging problems of NILM at Watty being one of the best disaggregation companies around the world
Software Engineer & Data Scientist
Perzonalization
2015-12 – 2016-05
Istanbul, TURKEY
  • Implementing and optimizing recommender systems on dataset of an online shopping website
  • Analyzing different approaches on the performance of the recommender system including Deep Learning approaches
Mobile Application Developer (REMOTE)
MizPiz Inc.
2014-10 – 2015-06
London, CANADA
  • As an iOS developer I developed MizPiz app for iOS whole application from the beginning to the end. MizPiz is designed to revolutionize how restaurant goers interact with each other and with restaurant owners in the effort to create a tightly connected community who all share one passion: food. With one app, you can save money, find new restaurants, and share your dining experiences with your friends to help the community grow. MizPiz will make dining worry free, enjoyable, and more affordable!
Mobile Application Developer
Mobven
2014-08 – 2014-10
Istanbul, TURKEY
  • For a short period of time I had the chance to work in Mobven's great team and it's professional environment. I was working on an Android application named HesapAl
Mobile Application Developer
Dade Pardaz Sahand
2013-01 – 2013-12
Tabriz, IRAN
  • Android online bus ticket booking app
Software Developer
Nar-Co
2012-01 – 2012-12
Tabriz, IRAN
  • Worked on several small projects including Web Design, Mobile Application Development, and Windows .NET application programming
Software Developer
Tabriz University Robotic Group (TURG)
2009 – 2013
Tabriz, IRAN
  • Worked as a team member and also software developer head in University of Tabriz. Including several sections: Data Acquisition from UDP port, De-serializing data using google protobuf, Extended Kalman Filter for Noise Reduction, Graphical Visualization of Robots and Ball with OpenGL, Strategy Planner Module, ERRT for path planning, Sending Out Commands using USB port to a RF-Module