Keywords
Skills
Project history
Actively supporting the team along different steps of the V Model to bridge the gap between functional and
software development and ensure the project is guided towards a successful completion
• Utilized CadCAD methodology to create a simulation environment for the vault
• Integrated Deep Reinforcement Learning techniques to develop an adaptive policy for the vault’s actions
• Enabled the vault to learn from its interactions with the simulated environment and investors behaviors, improving its
performance over time
Full Stack Web & Mobile APP Development
Description: German freelancing platform that matches between doctors and hospitals
-
Brainstormed and defined the core features (functional & non-functional requirements) to address client’s needs
-
Put together and managed remotely an international team of 6 members (Scrum master, UI/UX designer, Frontend,Backend, Android & iOS mobile developer and DevOps engineer)
-
Planned and managed the app development following the agile scrum methodology
-
Designed the software architecture: databases, functions and system behavior
-
Optimized the cloud infrastructure’s costs while maintaining the app’s performance within KPIs
-
Established the CI/CD pipeline using containerization (Docker)
-
Designed and implemented a machine learning algorithm to optimize the matching function and thus enhance the user-experience
-
Held biweekly meetings with the client and daily alignments with the dev-team to ensure costs, time and quality are met
Deep Learning Engineer, Autonomous Driving
Description: Highway Pilot Situation Interpretation using Deep Learning techniques
-
Led the collaborative development of a domain classification concept (Highway/City) between different compa- nies: Astech, EFS and Audi
-
Designed the logical and detailed architecture of the common training pipeline for domain classification, corridor estimation and lane change manoeuver prediction modules
-
Took part in developing the training pipeline and debugging the models training process (CNN, LSTM, FFN)
-
Established the embedded realtime inference of the trained models
Machine learning based Advanced Driver Assistance Systems
Description: Machine Learning based turning recognition for KAS performance enhancement
• Analysed KAS system limitations causing False-Positives (false car braking actions)
• Defined a machine learning based concept for FP mitigation and introduced it to Audi’s innovation board • Managed a team of 5 persons to achieve a POC for improving KAS behavior
• Achieved a 75% reduction of KAS false actions rate
* Developed model based features conform to MISRA and functional safety norm ISO26262
* Optimized and validated iteratively the time to collision estimator Algorithm
* Introduced and implemented a Continuous Integration concept to the project. This allowed a fully automation
of our Build and release process to better meet our biweekly release deadlines
* Project configuration manager: version and sustainability management of all relevant configuration items
* Supported MIL, SIL and functional testing
Keywords: Model Based, MISRA, Functional Safety, Time To Collision, Kinematic Motion Models, CI, Configuration Plan
Tools: Matlab | TargetLink | C | PolySpace | Misra | dSPACE | PTC Integrity | Jenkins | DOORS | Python
Description: Robust Control of a Rotary Steering System to automate directional drilling operations
* Modeled and validated the dynamic behavior of a rotary steering system within its operating environment
* Developed a trajectory planning method based on the spline interpolation theory
* Designed a robust control system for efficient path tracking while drilling
* Supervised and coordinated 3 master students involved in the project
* Co-authored a patent: "Depth-based borehole trajectory control" - USA | Patent
Keywords: Nonlinear Systems, Stability Analysis, Fourier Transform, Model Predictive Control, Partial Differential Equations
Tools: Matlab | Simulink | Stateflow | Acado | TFS | BHAsysPro
Simulated the dynamic behaviour of an injection nozzle and optimizated its operating point by solving a Pareto
Optimality Problem
Keywords: Numerical modeling, Gradient Analysis, Convex Optimization, GUI
Tools: Matlab Optimization Toolbox | Simulink