About Me
Aishwarya Jauhari Sharma
Applied AI Engineer · Germany
Applied AI engineer with four years of full-time experience across applied research and consulting, building machine learning and LLM systems and putting them into production.
Professional Summary
Four years of full-time experience across applied research at Fraunhofer FHR and consulting at Accenture, plus a working-student analyst role at mobile.de. Recent focus is retrieval-augmented generation and production MLOps.
🌟 Core Expertise
🧠 GenAI & Deep Learning
- LLMs
- RAG Systems
- Agentic Document Workflows
- Document Parsing
- Vector Databases
- Transformers & CNNs
- Image Segmentation · OCR · Transfer Learning
- Autoencoders & GANs
📊 Data Science & Analytics
- Classification · Regression · Clustering
- Feature Engineering & Hyperparameter Tuning
- Churn Prediction & Time-Series Analysis
- Customer Journey & Product Analytics
- A/B Testing · Statistical Inference
- EDA & Outlier Detection
🚀 MLOps & Engineering
- Python · PyTorch
- Vertex AI · BigQuery
- REST APIs · Docker · Linux
- CI/CD Pipelines (GitHub Actions)
- MLflow · Git · MkDocs
🛠️ Tools & Platforms
- Jupyter Notebook · PyCharm
- Tableau · Google Data Studio · Streamlit
- Confluence · Jira
Professional Experience
Current Status
Available for new opportunities - Actively seeking Data Science/AI-ML Engineering roles
Applied Researcher, Machine Learning
- Deep learning for Synthetic Aperture Radar image classification, cutting manual target detection effort by 22%
- Production MLOps pipeline for SAR change detection on FastAPI and Google Cloud Run
- Built a RAG engine so the team could query research papers in natural language
Data Analyst (Working Student)
- 12% lift in lead conversions from user drop-off prediction models
- Cut daily marketing report generation from 8 minutes to 2 through SQL and database tuning
- Dashboard tracking 20+ campaigns with KPI monitoring
Business Consultant
- Risk assessment and k-means segmentation for a US mortgage client, 4% portfolio growth
- Product analytics dashboards across a Canadian retail client's 200+ categories
Education & Certifications
| Degree | Institution | Year | Focus |
|---|---|---|---|
| M.Sc. Digital Engineering | Otto-von-Guericke University, Germany | 2019-2022 | Data Science and ML |
| B.E. Instrumentation Technology | Visvesvaraya Technological University, India | 2012-2016 | Control Systems and Signal Processing |
Thesis: Deep Learning for Image Segmentation: U-Net with Residual Learning and Attention Mechanisms
Key Projects & Publications
Featured Projects
| Project | Technologies | Impact |
|---|---|---|
| MLOps Auto-Retraining Pipeline | MLflow, FastAPI, Docker | Production-grade churn prediction |
| Multi-Modal RAG Chatbot | pyMuPDF, Weaviate, LangChain | Advanced document processing |
| End-to-End PDF Extraction | LLMs, Vector Stores | JSON/Markdown conversion |
| Doctract: Document extraction and query tool | Mistral 7B, RAG Architecture | Document extraction & querying |
Research Publications
- Comparison of Deep Learning Algorithms for Semantic Segmentation in Medical Imaging
- Unsupervised Anomaly Detection in Brain MRIs using Compact Context-Encoding Variational Autoencoder
Key Strengths
🚀 Full-Stack AI | 📊 Analytical Thinking | 🤝 Stakeholder Management | 🌱 Adaptability | ⚡ Learning Agility | 👥 Collaboration
Last Updated: January 2025