Machine Learning & Data Science
- Model Development
- Computer Vision (OpenCV)
- NLP
- Time Series Forecasting
- Recommendation Systems
- TensorFlow
- PyTorch
- scikit-learn
- Keras
Open to new opportunities
Arian
Azmoudeh
I build machine learning systems that leave the notebook — multi-agent assistants, computer-vision pipelines and autonomous robots that hold up in production.
Currently
MSc Financial EngineeringWorldQuant University
- Based in
- Switzerland
- Focus
- Applied AI & ML systems
- PyTorch
- TensorFlow
- Computer Vision
- LangGraph
- RAG
- Kubernetes
- ROS2
- MLflow
- FastAPI
- GCP
- 10+
- Years in AI
- 6
- Companies
- 22
- Certifications
- 3
- Publications
01About
A decade turning research into products
I have spent ten years on the unglamorous half of machine learning — the part where a model has to survive contact with real users, real latency and real budgets.
That work has taken me from multi-agent chatbot architectures and retrieval pipelines to a painting drone's vision stack and a ground robot built from an empty URDF file. The through-line is the same each time: understand the domain deeply enough to know which 5% of the problem actually needs a model, then engineer the other 95% properly.
Alongside the engineering I have led product and cross-functional teams, moderated more than seven industry webinars on disruptive innovation, and published research on sentiment-driven market prediction. I am now completing an MSc in Financial Engineering at WorldQuant University, pulling that quantitative rigour back into the systems I build.
- Based in
- Switzerland
- Focus
- Applied AI & ML systems
- Languages
- English, Persian, Italian
Where I go deep
- 01
Multi-agent systems & RAG
Retrieval pipelines and agent graphs that stay coherent past the demo.
- 02
Computer vision
Detection and inspection stacks, from a painting drone to medical imaging.
- 03
MLOps & cloud delivery
Containers, CI/CD and tracking so a model ships more than once.
- 04
Robotics & navigation
ROS2, SLAM and Nav2 on hardware built from an empty URDF file.
- 05
Quantitative finance
Sentiment-driven prediction and the MSc rigour behind it.
02Experience
Six roles, four countries, one thread
Engineering, data science and product — from financial systems in Tehran to disruptive AI in Lugano.
2023—25
Machine Learning Engineer
Swiss Institute for Disruptive Innovation
Lugano, Switzerland
May 2023 — November 2025
- Developed AI solutions aligned with business goals and optimized multi-agent chatbots to improve accuracy and engagement.
- Built and integrated AI agents into workflow automation platforms, enabling adaptive, data-driven automation for companies.
- Collaborated on a robotic drone system (PaintX) to streamline painting processes via advanced computer vision and robotics.
- Contributed to an AI Recruitment Platform enabling automated candidate sourcing and screening, reducing hiring time.
- Migrated the existing platform and AI solutions to a scalable cloud environment using DevOps best practices.
- Moderated and presented 7+ professional webinars and events focused on disruptive innovation and advanced AI technologies.
2021—23
ML Engineer & Product Manager
Olomax
Ticino, Switzerland
February 2021 — March 2023
- Led the design and creation of the Holographic Assistant for banking and healthcare, elevating user engagement and experience.
- Implemented advanced computer vision and NLP pipelines for emotion recognition and context-aware interactions.
- Improved conversational systems using RAG and NLP methodologies, enhancing accuracy and adaptability.
- Implemented cloud and DevOps strategies for secure, scalable AI deployment, optimizing containerization and CI/CD pipelines.
2019—20
Data Scientist & Project Manager
Mars Planet Technologies
Bergamo, Italy
January 2019 — December 2020
- Developed and managed chatbot solutions tailored to space-related industries, helping companies explore new market opportunities.
- Led cross-functional teams to implement AI-driven data analysis tools, improving operational efficiency in the space sector.
- Applied statistical and predictive modeling techniques to analyze data and derive actionable insights for strategic decisions.
- Enhanced data quality through advanced feature engineering, data preprocessing and data augmentation techniques.
2016—19
Full Stack Developer & DevOps Engineer
AriaHelpDesk
Dubai, United Arab Emirates
September 2016 — January 2019
- Developed and maintained full-stack web applications, integrating front-end interfaces with back-end services.
- Automated deployment pipelines using Docker and Kubernetes, reducing release cycles and enhancing reliability.
- Optimized APIs and database queries to boost performance, supporting large-scale traffic.
- Investigated and integrated emerging ML algorithms, applying predictive modeling for actionable insights.
2014—16
Data Scientist & Marketing Manager
Teb Gostar
Istanbul, Turkey
April 2014 — June 2016
- Analyzed monthly active user data to inform and optimize marketing and product strategies.
- Led the design and implementation of training and mentoring programs, reducing onboarding time for new hires by 50%.
- Achieved a 2× increase in average app engagement time and reduced drop-off rates by 30% through targeted data analysis.
- Identified and implemented strategies to reduce return rates by 10%, leading to significant cost savings.
2012—14
Full Stack Developer
Padide Paydar Nikan
March 2012 — February 2014
- Developed and maintained financial payment, statement and inventory systems, ensuring high performance and reliability.
- Optimized software performance through code refactoring and debugging, enhancing system efficiency.
- Conducted software testing and validation to ensure robust, error-free deployments.
03Toolkit
The stack I reach for
Grouped by the kind of problem it solves, not by vendor.
Agentic AI & LLM Integration
- RAG
- LangChain
- LangGraph
- LlamaIndex
- AutoGen
- CrewAI
- Copilot Studio
- Google AI Studio
- n8n
- Make
MLOps, DevOps & Cloud
- Docker
- Kubernetes
- GitHub Actions
- Jenkins
- Kubeflow
- MLflow
- AWS
- GCP
- Azure
- Terraform
- Ansible
- Linux
Programming & Frameworks
- Python
- C++
- C#
- Go
- Java
- PHP
- TypeScript
- FastAPI
- Django
- Flask
- Spring Boot
- .NET
- Next.js
- React Native
Data & Infrastructure
- PostgreSQL
- MySQL
- MS SQL Server
- MongoDB
- Neo4j
- Snowflake
- Pinecone
- Faiss
- Chroma
Analysis & Visualization
- Pandas
- NumPy
- Matplotlib
- Seaborn
- Plotly
- Tableau
- Power BI
- A/B Testing
- EDA
Robotics & Navigation
- ROS2
- URDF
- SLAM
- Navigation2
- Raspberry Pi
- Jetson Nano
- SolidWorks
Product & Leadership
- Agile
- Scrum
- Jira
- ClickUp
- Stakeholder Engagement
- Strategic Planning
- SEO/SEM
- Google Analytics
- HubSpot
04Selected work
Systems I designed or led
From holographic assistants for banking to an autonomous ground robot built from an empty URDF file.
01Human interface
Holographic Avatar
A 3D interactive avatar driven by real-time AI, combining natural language understanding with realistic facial animation for banking and healthcare.
- NLP
- Computer Vision
- Real-time 3D
02Robotics
Mobile Robot Platform
An autonomous ground robot built from scratch — URDF model in ROS2, LiDAR and camera fusion, SLAM mapping and Nav2 path planning.
- ROS2
- SLAM
- Gazebo
- Nav2
03HR technology
AI Hiring Platform
Recruitment suite using predictive analytics and skill-matching to automate candidate sourcing and screening at enterprise scale.
- Predictive Analytics
- NLP
- Ranking
04Healthcare
AI Medical Platform
Diagnostic system applying computer vision and deep learning to medical image analysis and patient prognosis.
- Deep Learning
- Medical Imaging
- MLOps
05Education
AI Learning Platform
Adaptive learning ecosystem generating personalized paths and content, tuned to each learner's pace and gaps.
- LLM
- Personalization
- RAG
06Growth
AI Marketing Platform
Marketing automation for predictive customer-behaviour analysis, segmentation and cross-channel campaign optimization.
- Big Data
- Segmentation
- Forecasting
05Research
Published, reviewed, in progress
Work on sentiment-driven market prediction, the limits of LLM evaluation, and object detection.
- 02
Leveraging LLMs to Navigate and Address Limits in Sentiment Analysis Models
Under review - 03
Advanced Object Detection Using Deep Learning Techniques
In progress
06Credentials
Three degrees, 22 certifications
Every certificate links straight to its issuer — nothing here is unverifiable.
- 01In progress
MSc Financial Engineering
WorldQuant University
January 2025 — Present
- 02
MSc Computer Science — AI & Machine Learning
Contemporary Technology University
January 2022 — June 2023
- 03
BSc Computer Science — Software Engineering
Shamsipour Technical and Vocational Academy
September 2007 — May 2013
Certifications
22 total07Gallery
Off the clock
Conferences, mountains and the occasional suit.
08Contact
Let's build something that ships.
Based in Switzerland and open to strategic partnerships and high-impact AI work — whether you have a defined problem or just a direction worth exploring. LinkedIn is the fastest way to reach me.
Message me on LinkedIn