Open to full-time roles · Berlin · open to EU relocation

Adhish
Nanda

I build data and AI systems end to end: analytics engineering pipelines, retrieval and ranking models, and dashboards built to hold up in production.

MSc Data Science in Berlin, around 2.5 years in enterprise analytics, and five peer-reviewed publications in IEEE and Springer.

Data Engineer Analytics Engineer Applied AI / ML Engineer Data Analyst / BI Junior Data Scientist
Download CV ↓ LinkedIn → GitHub ↗ Email
About

Who I am

I am a data and analytics professional based in Berlin, finishing an MSc in Data Science at the University of Europe for Applied Sciences. Before the master's I spent around 2.5 years in enterprise analytics in India, where I designed machine learning case studies, built Power BI dashboards, and shaped data solutions for large corporate clients.

I like problems that run from raw data all the way to something a person can use. In practice that means modelling the data properly, testing it, and making the result clear to people who do not read code. My recent work covers analytics engineering with dbt and BigQuery, a production-inspired ranking system, and an MSc thesis on retrieval and cost-aware LLM routing for AI governance.

I work in the open. Most of my projects live on GitHub with tests, documentation, and live demos, and several are backed by peer-reviewed research. I use AI coding assistants day to day, and I can explain, debug, and reason about the design decisions behind everything I ship.

I am available immediately and open to relocation across the EU. On graduation my student permit converts to an 18-month post-study job-seeker permit, so no employer sponsorship is required, and I am eligible for the EU Blue Card on a qualifying offer. German is at A2 and improving.

Adhish Nanda Berlin, Germany
// core toolkit
PythonSQLdbt CoreBigQueryPower BITableauLooker StudioPyTorchscikit-learnLightGBMFastAPIDockerRedisDuckDBFAISSBM25RAGASAirflowStreamlitGitGitHub ActionsPandas
GitHub activity summary for adhishnanda
Verified results, not just skills
0.97
Context precision on the thesis gold evaluation set of 25 annotated queries
72.2%
Generation cost reduction against an always-expensive routing baseline, at 109.4% quality retention
5
Peer-reviewed publications in IEEE and Springer venues
8
End-to-end projects on GitHub, several with tests, CI, and live demos
Selected work

Projects

A curated set of things I have built and shipped, from analytics engineering to retrieval and ranking. Most include tests, documentation, and a live demo.

MSc Thesis · Completed 2026
RAGLLM RoutingFastAPIDockerFAISS + BM25RAGASEU AI Act

Cost-Aware, Governance-First RAG System with Adaptive LLM Routing

A production-inspired retrieval system for AI-governance question answering across the EU AI Act, GDPR, NIST AI RMF, and ISO/IEC 42001. Hybrid retrieval combines symbolic exact-match, BM25 and FAISS fusion, and cross-encoder reranking, built from scratch. A rule-based router then sends each query to a cheap or costly model tier, evaluated with RAGAS and a counterfactual comparison against an oracle.

0.97
Context precision
(25-query gold set)
72.2%
Generation cost
reduction
109.4%
Quality retained
vs baseline
Apr 2026 – Aug 2026 · UE Berlin · Supervisors: Prof. Dr. Iftikhar Ahmed, Prof. Dr. Talha Ali Khan
Read case study → GitHub ↗
Live dashboard
dbt Core 1.11BigQueryKimball Star SchemaGitHub Actions CILooker StudioSQL

End-to-End dbt Analytics Stack on E-Commerce Data

An analytics engineering stack on 100,000-plus real Olist e-commerce orders. A three-layer dbt pipeline implements a Kimball star schema in BigQuery, with 94 automated tests run in GitHub Actions CI and a live Looker Studio dashboard covering revenue, delivery, geography, and product categories.

15
dbt models
across 3 layers
94
Automated
data tests
91.9%
On-time delivery
insight
May 2026 · Analytics Engineering
Read case study → Dashboard ↗ GitHub ↗
Live dashboard
DuckDBApache AirflowGitHub ActionsStreamlitPythonETL

German Job Market Analytics

A data-engineering pipeline that tracks the German data job market from source to dashboard. It ingests postings from three live sources, the Bundesagentur für Arbeit API, StepStone, and LinkedIn, then deduplicates and canonicalises them into a DuckDB warehouse of 574 records. Indeed was excluded after it blocked automated access.

An Airflow DAG runs a daily refresh through GitHub Actions, and 347 automated tests guard data quality. A live Streamlit dashboard surfaces roles, skills, and hiring trends. Built on the market I am applying into.

3
Live ingestion
sources
574
Canonical
job records
347
Automated
data tests
2026 · Data Engineering
Live dashboard → GitHub →
Live demo
LightGBMFastAPIRedisDuckDBStreamlitIPS / SNIPS

Personalised News Feed Ranking System

A full ranking pipeline built from scratch: article ingestion into DuckDB, point-in-time feature engineering with no data leakage, a LightGBM ranker with a logistic-regression baseline, and multi-objective reranking for freshness and diversity.

A Redis online feature store with graceful fallback serves features at low latency behind a FastAPI service. Offline evaluation uses IPS and SNIPS counterfactual estimation, so policy improvement is measured without an A/B test. A Streamlit UI exposes the live feed and health checks.

Mar 2026 – May 2026 · 43-test suite, full pass rate
Live demo → GitHub →

More projects

PyTorchDistilBERTCNN-BiLSTMResNet

Multimodal Emotion Recognition

Late-fusion emotion recognition across text, audio, and video using frozen pretrained encoders and a trainable fusion head. Reached 54.1% accuracy and 40.8% macro-F1 on the MELD benchmark. MSc capstone at UE Berlin.

Oct 2025 – Jan 2026 GitHub →
MediaPipescikit-learnComputer Vision

Gesture-Controlled German Learning App

Real-time gesture recognition for an interactive language-learning app. Pose detection and landmark extraction through MediaPipe, gesture classification in scikit-learn, and a full pipeline from data collection to prediction.

Nov 2025 – Jan 2026 GitHub →
Apache AirflowPythonETLStreamlit

Airflow-Orchestrated Polling Pipeline

A batch data-engineering pipeline that automates ingestion, cleaning, feature engineering, and batch prediction on polling data through modular Airflow DAGs, with a Streamlit dashboard for trends. An academic project.

Nov 2024 – Jan 2025 · Academic project GitHub →
Random ForestXGBoostK-MeansIsolation Forest

Healthcare Cost Prediction & Risk Segmentation

An end-to-end ML project for insurance cost prediction, comparing Linear Regression, Random Forest, and XGBoost. K-Means segments customers and Isolation Forest flags unusual high-cost cases. A course project at UE Berlin.

Dec 2024 – Jan 2025 · Academic project GitHub →
Experience

Where I have worked

Sep 2024 – Present
MSc Data Science Student
University of Europe for Applied Sciences · Berlin, Germany
PyTorchdbtBigQueryFastAPIAirflowPythonSQLDocker
  • Built end-to-end ML and analytics systems alongside coursework, including a dbt stack on BigQuery with 94 automated tests and CI, and a ranking pipeline with a live demo
  • MSc thesis on cost-aware, governance-first retrieval-augmented generation, cutting generation cost by 72.2% while retaining 109.4% of baseline quality
  • Capstone on multimodal emotion recognition across text, audio, and video, reaching 54.1% accuracy on the MELD benchmark
  • Core modules: Machine Learning, Data Engineering, Cloud Computing (AWS), NLP, Data Visualisation, Decision Support Systems
Feb 2022 – Sep 2024
Solutions Design Lead, Data & Analytics
NIIT Ltd (StackRoute) · Bangalore, India
PythonSQLPower BITableauscikit-learnPandasETL
  • Designed 20-plus machine learning case studies across regression, classification, clustering, and NLP for enterprise learning programmes
  • Built 6 to 7 Power BI dashboards and SQL reporting for programme KPIs and executive visibility
  • Wrote Python scripts for data transformation, validation, and reporting, reducing manual work and standardising data quality
  • Contributed to presales solutioning: feasibility analyses, stack recommendations, and solution design used in client proposals
  • Acted as the data and analytics point of contact for 40 to 50 enterprise clients and supported 20 to 30 programme staff
Jul 2021 – Jan 2022
Business Intelligence / Data Analyst Intern
NIIT Ltd (Learning Science & Technology) · Gurugram, India
Power BIPythonSQLETL
  • Built interactive Power BI dashboards for KPI transparency and executive reporting across divisions
  • Automated recurring reporting workflows with Python, removing manual steps from reporting cycles
  • Validated programme metrics with SQL and maintained documentation for audit readiness and handover
Mar 2021 – May 2021
Marketing Data Analyst Intern
Weaddo · Gurugram, India
PythonTableauRegressionFunnel Analytics
  • Used regression and funnel analysis on multi-channel marketing data to find conversion drivers
  • Tracked ROAS, CPA, CAC, and conversion rate in Python, and built Tableau dashboards for live campaign monitoring
May 2019 – Jul 2019
Data Research Analyst Intern
Amity Institute of Integrative Science and Health · Gurugram, India
PythonCNNSQLRTableau
  • Built a CNN image-classification pipeline for breast cancer detection, contributing to a Springer-published paper
  • Cleaned and processed clinical datasets in Python, SQL, and R, handling class imbalance and normalisation
Jan 2017 – Jun 2017
Academic Researcher
Anandalal Electric Private Limited · Bengaluru, India
PythonTime-SeriesRegression
  • Built a time-series regression model for Li-Ion battery state-of-charge estimation, published in AJAST 2017
  • Ran load-forecasting analysis for electric vehicle charging contexts
Education

Academic background

MSc Data Science · Expected Aug 2026
University of Europe for Applied Sciences, Berlin
Sep 2024 – Aug 2026
Machine LearningData EngineeringCloud Computing (AWS)NLPData VisualisationDecision Support Systems

Thesis (completed): Design and Evaluation of a Cost-Aware, Governance-First Retrieval-Augmented Generation System with Adaptive LLM Routing and Counterfactual Evaluation. Hybrid retrieval, adaptive routing, and RAGAS evaluation across four AI-governance corpora. Supervised by Prof. Dr. Iftikhar Ahmed and Prof. Dr. Talha Ali Khan.

Capstone: Multimodal emotion recognition fusing text, audio, and video. 54.1% accuracy and 40.8% macro-F1 on the MELD benchmark.

Postgraduate Summer Programme · 2019
Ludwig-Maximilians-Universität München (LMU), Germany
Aug 2019 – Sep 2019
Supervised MLUnsupervised MLClusteringPrediction

Applied machine learning and analytics to financial datasets at one of Germany's leading research universities. Early exposure to academic research methodology in an international setting.

MTech Data Science · 2018–2020
Amity University, Gurugram, India
Aug 2018 – Jun 2020
Data MiningPredictive AnalyticsNLPBig DataBayesian NetworksOLAP

Thesis: Transaction Scheduling using Machine Learning. Modelled scheduling strategies with supervised learning and structured feature engineering. Research output: two IEEE publications.

BE Electronics & Communication · 2013–2017
MVJ College of Engineering (VTU), Bengaluru, India
Aug 2013 – Jun 2017
Engineering MathematicsSignals & SystemsDSPC / C++

Graduated First Class. Built the mathematical foundations, linear algebra, probability, and signal theory, that underpin current ML work. Final-year project on Li-Ion battery state-of-charge estimation, published in AJAST 2017.

Research

Publications

Five peer-reviewed papers across IEEE and Springer venues, spanning machine learning, OLAP, medical imaging, and applied electronics.

IEEE
A Comprehensive Survey of Machine Learning in Scheduling of Transactions
IEEE ICOEI · Jul 2020 DOI →
IEEE
A Comprehensive Survey of OLAP: Recent Trends
IEEE · Dec 2019 View →
Springer
Analysis of Breast Cancer Detection Techniques Using RapidMiner
SpringerLink · Jul 2020 View →
AJAST
SOC Estimation of Lithium-Ion Batteries used in Electric Vehicles
Asian Journal of Applied Science and Technology · Jun 2017 PDF →
AJAST
Implementing a Smart-Vehicle using IoT and Cloud Computing Extension
Asian Journal of Applied Science and Technology · Jun 2017 PDF →
Credentials

Certifications

DeepLearning.AI TensorFlow Developer
DeepLearning.AI / Coursera · Nov 2025
Verify →
IBM Data Science Professional Certificate
IBM / Coursera · Jan 2022
Verify →
A-Z Hands-On Tableau Training for Data Science
Udemy · Jan 2022
Verify →
Microsoft Power BI Desktop for Business Intelligence
Udemy · Aug 2021
Verify →

Open to relocation across the EU

Available for full-time roles, immediate start possible. English-first environments preferred.

Germany (base) Netherlands Ireland Luxembourg Austria Belgium Denmark
Contact

Let us talk about a role

Open to full-time roles in data engineering, analytics engineering, applied AI and ML, and data or BI analysis, across Germany and the EU.

Work authorisation available · EU Blue Card eligible · English fluent · German A2 and improving

Download CV ↓ Email LinkedIn ↗ GitHub ↗ +49 162 4758595