praneeth.

AI & DATA SCIENCE · THE NETHERLANDS

Praneeth
Dathu.

I like understanding how things work, then trying to build them myself. AI has given me plenty to work on — and plenty more to learn.

I’m an EngD Trainee in Data Science at JADS, Den Bosch.

See my projects Research profile
Currently learning more about deploying AI systems.
Praneeth Dathu wearing a navy jacket and glasses
Praneeth DathuMSc Artificial Intelligence · Leiden
Just started at
JADS, Den BoschSEP
2026
01 / ABOUTWHAT KEEPS ME CURIOUS

I learn most
by building.

Reading gives me a place to start. Building is where I begin to understand things: trying an idea, seeing what works, and figuring out what to change.

During my MSc in AI at Leiden, I worked with Bluetooth sensor data at Almende to estimate where something is indoors. I also built with language models and explored how agents use memory. Each project gave me something new to learn.

My experience and education
01

Working with real data

My thesis involved collecting sensor readings, training models and checking what their predictions could tell us.

02

Trying ideas with LLMs

I’ve been exploring conversational memory, retrieval and connecting language models to other tools.

03

Learning what comes next

I want to get better at deploying what I build — using cloud services, monitoring models and keeping systems running.

Mariënburg entrance at JADS in Den Bosch, from my first-week announcementMARIËNBURG · ’S-HERTOGENBOSCH

NOW / SEPTEMBER 2026

I’ve just
joined JADS.

In September 2026, I started my EngD in Data Science at JADS in Den Bosch.

I want to understand more of what it takes to put AI into use. I’m looking forward to working on industry problems, learning from the people around me and building on what I’ve done so far.

EngD Data ScienceSeptember 2026 — present
Read my first-week update
02 / SELECTED WORKTHESIS, COURSEWORK & PROJECTS

What I’ve
been building.

Some projects from my studies and work,
plus ideas I wanted to try for myself.

BLE / INDOOR LOCALISATION01
Conceptual diagram of Bluetooth anchors locating a point across an indoor floor plan estimated position
BLE anchor Position estimateConcept illustration

MSC THESIS · ALMENDE / LEIDEN

3D indoor localisation with Bluetooth

For my MSc thesis, I used Bluetooth signal strength to estimate indoor positions. I worked on the pipeline from collecting sensor data to running predictions in a live system.

PythonPyTorchBLEMachine learning
More about this project

I collected data from a Crownstone mesh and compared classical machine learning with deep learning. A large part of the work was deciding how to evaluate the models and what the results actually showed.

I also worked on preprocessing, repeatable experiments and real-time inference. It gave me experience with the steps around a model, as well as the model itself.

Project update on LinkedIn ↗

LLM SYSTEMS · RESEARCH

A conversational agent with memory

I worked on an agent with a defined persona and retrieval-based memory, so earlier conversations could inform its replies.

RAGVector retrievalPrompt caching
More about this project

I put together the persona prompts, conversation flow and memory retrieval, then deployed the agent on a VPS. I also explored prompt caching and model choices to manage the cost of running it.

PERSONAL PROJECT · VAULT AI

Vault AI: a personal finance app

I built Vault AI to organise transactions and make spending easier to understand, with a language model helping explain the patterns.

PythonSQLiteLLM API
More about this project

Transactions are stored in SQLite. The app groups spending, highlights recurring costs and unusual activity, and supports backup and sync through Google Drive.

APPLIED AI · ALMENDE

An AI assistant for internal tools

At Almende, I worked on an internal dashboard that connects a chatbot to tools and data through MCP.

GroqMCPDashboard
More about this project

I used Groq for language-model responses and MCP to connect the assistant to tools and data. I built the dashboard interface with Lovable.

GAME AI · LEIDEN

A Pacman capture-the-flag agent

Our team built a Pacman agent using Monte Carlo Tree Search. We placed fourth out of 37 teams in the course competition.

PythonMCTSMulti-agent systems
More about this project

I worked on combining search with game-specific heuristics. We used match simulations to test how the agent balanced attacking, defending and reacting to opponents.

RECOMMENDER SYSTEMS

Movie recommendations with neural networks

I built a recommender using MovieLens 1M to explore how a model learns from user ratings and ranks films they might like.

PyTorchMovieLensRanking
More about this project

I implemented neural collaborative filtering in PyTorch, trained it on an RTX 3090 and evaluated the ranked recommendations.

You can find more of my work on GitHub.Visit my GitHub
03 / EXPERIENCE & EDUCATIONSO FAR

My experience
& education.

I studied computer science, then AI.
Now I’m continuing with an EngD
in Data Science at JADS.

My research profile ↗
  1. SEP 2026 — PRESENT CURRENT

    EngD Trainee in Data Science

    JADS · Den Bosch

    I’m continuing my training in data science, with an interest in how AI is developed and used in industry.

    JADSTU/e
  2. COMPLETED AUGUST 2026

    MSc Artificial Intelligence

    Leiden University

    I completed my MSc with a thesis on BLE-based 3D indoor localisation. My thesis received an 8/10.

    Leiden University
  3. 2026

    Machine Learning Research Intern

    Almende · Rotterdam

    I worked on my thesis here, collecting Bluetooth sensor data, comparing models and connecting predictions to a live system.

    Almende
  4. 2023 — 2024

    AI Researcher

    Digital Ellanky · Hyderabad

    I used Python and SQL to analyse marketing data, build predictive models and prepare reports.

  5. FOUNDATION

    B.Tech Computer Science Engineering

    Malla Reddy Engineering College · India

    My undergraduate degree in computer science and engineering.

04 / TOOLKITUSED ACROSS THESE PROJECTS

Tools I use.

These are tools I’ve worked with.
Cloud deployment and MLOps are areas I’m learning more about.

Models & experiments

Python · PyTorch · scikit-learn
CatBoost · NumPy · Pandas

Language & retrieval

LLM APIs · RAG · Vector retrieval
MCP · Prompt engineering

Pipelines & deployment

SQL · MQTT · InfluxDB · FastAPI
Docker · Linux · Git · SLURM

05 / NOTES & MILESTONESRECENT UPDATES

A few updates.

Visit the blog archive ↗

SEPTEMBER 2026 · JADS

Starting my EngD at JADS

A few photos from my first week in Den Bosch.

Read on LinkedIn

AUGUST 2026 · LEIDEN

Finishing my MSc in AI

I defended my thesis on indoor localisation and completed my MSc at Leiden.

Visit my LinkedIn

RESEARCH · ALMENDE

My indoor localisation project

An update on using Bluetooth signals to estimate positions indoors.

Read on LinkedIn

06 / CONTACT

Get in touch.

LinkedIn GitHub Research profile