About
About me
Data Scientist turned AI Engineer building end-to-end LLM-powered applications, RAG pipelines, and multi-agent systems.
Who I am
I'm a Data Scientist turned AI Engineer based in Berlin with over 7 years of experience applying ML, statistical modelling, and Generative AI to real-world business problems. I specialize in building end-to-end LLM-powered applications — RAG pipelines, agentic workflows, and multi-agent systems.
My sweet spot is transforming complex data into practical, intelligent systems. I work across the stack — data ingestion, modeling, evaluation, and deployment — using tools like Python, PyTorch, LangChain, and Hugging Face to create solutions that are both accurate and efficient.
I thrive on bridging classical data science with cutting-edge Generative AI, whether it's forecasting demand, classifying data, or shipping multi-agent systems. I have a strong track record of translating analytical outputs into actionable stakeholder insights.
Toolkit
Technical skills & proficiencies
A comprehensive overview of my technical expertise across various AI and development domains.
Languages & Frameworks
10GenAI & LLMs
8Backend & Infra
10Career
Professional journey
My career path and key milestones in the field of AI and machine learning.
Data Scientist
Flix Mobility Tech GmbH — BerlinJan 2023 – Present
Led development of a team-wide AI coding harness (bfr-knowledge) adopted by 5 data scientists. Engineered seat-type and vehicle-type features for tree-based demand models, redesigning the ML pipeline to support multi-seat pricing and delivering a 7.2% uplift in accuracy.
Data Scientist
Pixsy — BerlinJun 2021 – Jun 2022
Built an end-to-end ML pipeline using Apache Kafka, Spark Streaming, and Airflow for real-time image similarity at scale. Developed a hybrid model (ORB + ResNet/Siamese networks) that improved matching accuracy by 22% and reduced false positives by 12%.
M.Sc. in Data Science
Otto-von-Guericke Universität — MagdeburgOct 2019 – Dec 2022
Key Courses: Machine Learning, Deep Learning, Generative Models, Computer Vision, Recommender Systems, Data-Warehouse Technologies.
ML Engineer
Accenture AI — BengaluruNov 2018 – Aug 2019
Developed failure prediction models using Gaussian Mixture Models for trade transaction risk assessment (Recall: 0.768, F1: 0.704). Built a batch runtime prediction tool using Multivariate Linear Regression for proactive scheduling decisions.
Application Analyst
Accenture — Pune, IndiaJul 2016 – Nov 2018
Built an email classification and auto-ticket assignment system using Naïve Bayes and Linear SVM (F1: 0.797). Optimised SQL queries, reducing response time by 38%, and delivered EDA and visualisations on retail sales data.
B.Tech in Electronics & Communication
GGSIPU — Delhi, IndiaAug 2012 – May 2016
Key Subjects: Data Structures and Algorithms, C++, Software Engineering, Applied Mathematics and Statistics.
Principles
My approach
The core principles that guide my work.
Full-Stack Mindset
Building end-to-end solutions, from robust data pipelines and backend logic to intuitive user-facing interfaces.
Pragmatic Innovation
Applying the right tool for the job, whether it is a classic ML model or a fine-tuned LLM, to deliver efficient and scalable results.
Continuous Learning
Staying hands-on with the latest tools and techniques in the fast-paced world of AI to build cutting-edge, optimized systems.