Skip to content

Courses, theses and first research steps

Teaching & Mentoring

I teach and mentor the way a good plan works: start from where students are, set a clear goal, and make every step one they can take on their own. All course materials are open for anyone to learn from.

Office hoursMon 12:00 to 14:00, Wed 08:30 to 10:30, Room 104 Details

Six pawns advancing up a chessboard at different speeds; one has reached the last rank and become a queenabcdef12345678=Q
Every student moves at their own pace. The goal is that each one keeps moving.

Courses

UMT, bachelor, 14 weeks

Object-Oriented Programming in Java

Learn to think in objects. One application, built in three sprints and public on GitHub from week one.

Object modellingInterfaces & polymorphismRecords & streamsLambdasJUnit testingGit workflowDesktop GUI

Open the course

UMT, bachelor year 2

Data Structures & Algorithms

Programs that stay correct and fast as the input grows. Slides and practice for every lecture, and a three-sprint project.

Big-O thinkingHash maps in practiceRecursion & backtrackingTrees & BSTsGraph search: BFS, Dijkstra, A*Problem solving with AI

Open the course

UMT, bachelor year 3, 14 weeks

Machine Learning

Building things that work on real data. Three team projects, and a system you defend in front of the room.

End-to-end pipelinesLeakage & honest evaluationscikit-learnTree ensemblesPCA & clusteringAI-assisted workflows

Open the courseHolberton ML Handbook

Bachelor or master

Your thesis

The next move: one project of your own, in AI agents, agentic retrieval or robotics, supervised step by step.

Your questionYour codePossibly a paper

See the topics

How I teach

A learning loop of five steps, first principles, ask why, build, defend and share, around understanding, with AI as a companion ring around the whole loopunderstanding1First principles2Ask why3Build4Defend5ShareAI as a companion
  1. First principlesStart from the idea, not the syntax or the tool.
  2. Ask whyTrace every algorithm by hand until it makes sense.
  3. BuildOne real project that grows across the whole semester.
  4. DefendExplain your choices. A result you can explain beats a score that flatters you.
  5. ShareWork is public on GitHub, and explaining it to others locks it in.

AI as a companion. It sits around every step, never inside it: it explains, reviews and challenges, but the thinking stays yours.

Thesis topics

AI agentsBSc or MSc

An AI agent that plans with a motion planner

An LLM agent turns a task in plain language into goals and constraints, calls a sampling-based motion planner as a tool, and checks the path it gets back.

PythonLLM tool callingMotion planning

Agentic RAGBSc or MSc

An agentic RAG assistant that cites its sources

An assistant for course material that decides when to search again, answers with citations students can check, and is measured on how often it is right.

PythonEmbeddingsEvaluation

RoboticsMSc

Safety zones around moving people, in simulation

Extend safety-zone path planning so the margins adapt to how people move, and compare it with a fixed-margin planner in ROS 2.

ROS 2GazeboRisk-aware planning

Any area

Your own idea?

Bring a question you care about. If it fits the time you have and something I can supervise well, we shape it together.

Write to me

How a thesis works

  1. Pick a topicFrom the list, or your own idea
  2. One-page planQuestion, approach, what "done" means
  3. Build and testShort meetings every two weeks
  4. Write it upReviewed chapter by chapter
  5. DefendWith a rehearsal first

What I expectSteady work, honest updates when you are stuck, and code you can explain line by line.

What you can expectFocused meetings, feedback within a week, and help turning strong work into a paper.

Office hours and email

Office hours

Monday
12:00 to 14:00
Wednesday
08:30 to 10:30
Where
Room 104, UMT. No appointment needed

Emailing me

To
evis.plaku@umt.edu.al, from your university address
Subject
Course first, DSA: assignment 2, or Thesis: and the topic title
Include
Your group or program, the question, and what you tried

Replies usually within two working days. For a recommendation letter, ask three weeks ahead and send your CV and the program you are applying to.