CV
Downloadable CV and structured resume.
Contact Information
| Name | Muhammad Ammar |
| Professional Title | AI Research at Siemens | MSc Electromobility-ACES @ FAU |
| muhammad.ammar@fau.de | |
| Location | Erlangen, Bavaria |
| Website | https://ammar2k.github.io |
Professional Summary
Machine Learning engineer working on LLM pipelines, knowledge graphs, ontology-driven graph-based analytics, and applied AI. Pursuing an MSc in Electromobility-ACES at Friedrich-Alexander-Universitat Erlangen-Nurnberg while researching GenAI-enabled software ecosystem health monitoring solutions at Siemens.
Experience
-
2026 - Present Erlangen
Master's Thesis
Siemens AG + FAPS FAU
Beyond the Agent: A Workflow-Level Review of Agentic AI Systems.
- Research objective: How are agentic AI systems structured, controlled, and evaluated as end-to-end workflows, and where do current approaches lack system-level rigor?
-
2024 - Present Erlangen
Data Analytics Researcher - Working Student
Siemens AG
Research on software ecosystem health monitoring, ontology-driven ETL, and LLM-based knowledge graph analytics.
- Conducted a systematic literature review of software ecosystem health analysis, surveying 61 research articles and 416 health metrics to build a unified metric ontology for ecosystem health monitoring.
- Designed ontology-driven ETL pipelines that transform research metrics, KPIs, and platform ecosystem data into structured Neo4j knowledge graphs.
- Developed an LLM-based Text-to-Cypher pipeline that maps natural-language prompts to knowledge graph queries for software ecosystem health analytics.
- Modeled software ecosystem health as a metric knowledge graph, connecting indicators, KPI trees, and graph algorithms for real-time analysis and visualization.
- Developed an LLM-powered Prompt-to-Visualization pipeline to display the most relevant dashboard for a user query, lowering TTI from 2 minutes to 5 seconds.
- Submitted 6 invention disclosures and authored 4 research papers.
-
2025 - 2026 Erlangen
Project Thesis
FAPS FAU
Architectural Paradigms of Multi-Agent Systems and Agentic AI in Intelligent Manufacturing Systems, A Systematic Literature Review.
- Research objective: How have agent architectures evolved in manufacturing systems, and what architectural design principles emerge from this evolution that are relevant for manufacturing control?
-
2024 - 2024 Remote
Machine Learning Engineer
NeuroOceans AI
Built and evaluated audiovisual deepfake detection models.
- Finetuned multiple audiovisual deepfake detection models and evaluated them on performance metrics.
- Led the entire pipeline development, from dataset collection and preprocessing to model training and evaluation.
-
2023 - 2023 Remote
Solution Analyst
Learners.ai
Built GPT-powered workflow automations and chatbots using LangChain and Pinecone.
- Used LangChain and Pinecone to create a GPT-powered RAG chatbot that takes a sales call transcript and generates a sales proposal.
- Developed a copywriter chatbot using LangChain Agents, tool calling, and a curated knowledge base.
- Conducted Data Import Analysis for data transfer across CRMs, using Excel, SQL, and Python.
- Created ad campaign effectiveness dashboards in HubSpot and Databox.
-
2021 - 2022 Islamabad, Pakistan
Undergraduate Researcher
Robotics and Intelligent Systems Engineering (RISE) Laboratory
Senior year research project focused on mobile robotics.
- Engaged in a senior year research project, providing bi-weekly presentations to a supervisory panel.
- Developed a control algorithm for a mobile robot using Arduino, ultrasonic sensors, IR sensors, and encoders.
- Successfully secured funding through IGNITE National Grassroots Research Initiative 2022 and obtained funding for AHFE Conference registration fees under the Pakistan and European Unions Horizon 2020 ENHANCE project.
Education
-
2024 - Present Erlangen, Germany
MSc
Friedrich-Alexander-Universitat Erlangen-Nurnberg (FAU)
Electromobility-ACES
- Current grade: 1.4
-
2025 - 2025 Pisa, Italy
Seasonal School
Scuola Superiore Sant'Anna (SSSA)
AI & Robotics in Extended Reality Seasonal School
-
2018 - 2022 Islamabad, Pakistan
B.Sc.
National University of Sciences and Technology (NUST)
Mechanical Engineering
Publications
-
2025 Towards Data-Driven Real-Time Performance Monitoring of Platform Ecosystems.
APSEC 2025 Software Engineering in Practice track
Conference paper on real-time performance monitoring for platform ecosystems.
-
2025 Data-Driven Health Monitoring of Platform Ecosystems: A Systematic Literature Review.
SDS 2025
Systematic literature review on data-driven health and performance monitoring for platform ecosystems.
-
2025 Prompt-to-Metric: LLMs and Graph Algorithms for Platform Ecosystem Health Monitoring.
WIPIEC
Combining LLMs with graph analytics to monitor platform ecosystems.
-
2025 Best Paper Award
Empirical study of schema injection strategies for LLM-based Text-to-Cypher query translation.
-
2022 A Chain-Driven Live Roller Mechanism for Loading and Unloading Packages on Autonomous Mobile Robots in Warehouses.
Applied Human Factors and Ergonomics (AHFE) Conference
Conference paper on an autonomous mobile robot mechanism for warehouse loading and unloading.
Skills
Languages
Certificates
- How to Get Into Robotics - University of Leeds (2024)
- Docker Essentials: A Developer Introduction - IBM Cognitive Class (2024)
- Open Source Software Development Methods - The Linux Foundation (2023)
- Generative AI with Large Language Models - DeepLearning.AI and AWS (2023)
- Building AI Products with OpenAI - Uplimit (2023)
- Natural Language Processing Specialization - DeepLearning.AI (2023)
- PyTorch for Deep Learning in 2023 - Zero to Mastery - Udemy (2023)
- Web Developer Bootcamp with Flask and Python in 2023 - Udemy (2023)
- Deep Neural Networks with PyTorch - IBM (2023)
- YOLOv7 YOLOv8: Deep Learning - Computer Vision 2023 - Udemy (2023)
- Computer Vision Bootcamp with Python - Udemy (2023)
- Machine Learning Specialization - DeepLearning.AI and Stanford University (2023)
- Unsupervised Learning, Recommender Systems, Reinforcement Learning - DeepLearning.AI and Stanford University (2023)
- Advanced Learning Algorithms - DeepLearning.AI and Stanford University (2022)
- Supervised Machine Learning: Regression and Classification - DeepLearning.AI and Stanford University (2022)
- IELTS 8.5 - British Council (2022)
- Crash Course on Python - Google (2022)
- Understanding Research Methods - University of London and SOAS University of London (2022)
- Introduction to TensorFlow for AI, ML, and Deep Learning - DeepLearning.AI (2022)
- Supply Chain Operations - Rutgers University (2022)
- Supply Chain Logistics - Rutgers University (2022)
- Robotics: Aerial Robotics - University of Pennsylvania (2022)
- Introduction to Mechanical Engineering Design and Manufacturing with Fusion 360 - Autodesk (2022)
- SQL for Data Science - UC Davis (2021)
- Getting and Cleaning Data - Johns Hopkins University (2021)
- Data Analysis with R Programming - Google (2021)
- The Data Scientist's Toolbox - Johns Hopkins University (2021)
Projects
-
Autonomous Mobile Robot
Designed and developed an autonomous mobile robot for warehouse operations, including obstacle avoidance and an attachable live-roller mechanism for automated loading and unloading.
- Robotics
- Arduino Programming
- Controls
- Design
-
Image2Audio app
Users upload an image, the app generates a caption for the image, creates a short story based on the caption, and then converts the story into audio format.
- OpenAI API
- HuggingFace
- LangChain
- Streamlit
- TTS
-
Q&A Chatbot on private data
Blog post on Retrieval Augmented Generation (RAG) using ChromaDB, GPT wrapped around LangChain, and private data.
- OpenAI API
- LangChain
- Vector Database
- Blogpost
-
Scene Classification
Finetuned ViT pre-trained on ImageNet for Kaggle Intel Classification Challenge.
- PyTorch
- Computer Vision
- Gradio
-
Podcast Summarizer
Converts podcast audio to text using Whisper, uses GPT-3.5 to generate a summary, and uses the Wikipedia API to display additional information.
- OpenAI API
- Whisper
- Audio2Text
- Streamlit
-
Transfer Learning with PyTorch
Scene classification by fine-tuning pre-trained Vision Transformers (ViT) guide.
- PyTorch
- Vision Transformer
- Blogpost