Erion Ahmetaj

CS @ Stevens Institute of Technology
Class of 2029

Erion
Ahmetaj

Computer Science sophomore minoring in Machine Learning. Last summer I built and shipped AI agents at Pfizer that took manual data work off analysts' plates. I'm looking for internships and co-ops in AI, data and software.

erion@stevens: ~
erion@stevens:~$ whoami
Erion Ahmetaj · CS '29 · minor in Machine Learning

erion@stevens:~$ cat experience.log
2026  Data Science Intern, Pfizer
2024– Tech Associate, Pawsome Vita
2024  Research Delegate, YLC Chengdu

erion@stevens:~$ ls ./stack
python  java  c  c++  sql  arm64
databricks-genie  copilot  git  linux

erion@stevens:~$ 
01 / about

I like building tools that save people real time.

I'm a sophomore at Stevens studying Computer Science with a minor in Machine Learning, and I hold a 3.8 GPA. My coursework covers both sides of the field: the math behind machine learning (linear algebra, probability, multivariable calculus) and the systems underneath it (data structures, algorithms, computer architecture, ARM64 assembly).

At Pfizer I learned that a useful AI agent starts with listening. I met one-on-one with data scientists and business stakeholders, scoped where automation would help most, then built and tested agents in Databricks and Microsoft Copilot until they were accurate enough to rely on.

Outside of school I'm the technology associate at Pawsome Vita, a pet-care business in Jersey City, where I set up and support the systems the team uses every day. I've also spent years fundraising for maternal and child healthcare in Kosovo. I speak English and Albanian, plus some Spanish.

02 / experience

Experience

Pfizer · New York, NY

May – Aug 2026
Data Science Intern
  • Gathered requirements in 1:1s with data scientists and business stakeholders, then built and deployed AI agents in Databricks and Microsoft Copilot that automated large-scale data categorization analysts had been doing by hand.
  • Developed the DQS AI agent as a Databricks Genie space on IQVIA's Data Query System, a global platform for selecting clinical trial sites and investigators. Stakeholders can now ask plain-English questions and get self-service access to quality data.
  • Raised agent accuracy and efficiency by writing and running manual test cases, then refining prompts, instructions and agent logic based on the results.
  • Delivered a portfolio categorization project using AI-assisted classification to organize portfolio data for downstream analysis.

Pawsome Vita · Jersey City, NJ

Sep 2024 – Present
Technology Associate & Daycare Attendant (part-time)
  • Implemented and support the daily-use systems for scheduling, customer management and internal workflows.
  • Set up and maintain core infrastructure, including security cameras, phones and on-site systems.
  • Research regional competitors to refine pricing and identify customer-driven service improvements.
  • Standardized care routines and keep operations compliant with safety and hygiene protocols.

Youth Leadership in China · Chengdu

Summer 2024
Research & Leadership Delegate
  • Selected for a competitive three-week international program on sustainable innovation and global leadership.
  • Researched sustainable energy adoption in urban environments and presented a thesis to an expert panel, earning commendation for analytical rigor.
  • Worked with international peers on solutions for renewable energy access in developing cities.
03 / projects

Projects & research

AI agent · Pfizer internship

DQS AI Agent

A Databricks Genie space built on IQVIA's Data Query System, the database Pfizer uses to identify, evaluate and select clinical trial sites and investigators. Instead of writing queries, stakeholders ask questions in plain English and get answers from quality data. I tuned it through rounds of manual test cases, adjusting prompts, instructions and logic after each pass.

Databricks GeniePrompt engineeringSQLAgent testing
AI classification · Pfizer internship

Portfolio Categorization

Applied AI-assisted classification to organize portfolio data for downstream analysis, replacing categorization that had been done manually.

Microsoft CopilotDatabricksData analysis
Game development

2-D Java Arcade Game

An interactive arcade game, more than 1,000 lines of Java Swing, built on an object-oriented architecture with collision detection. I debugged and optimized it for stable frame rates and responsive controls.

JavaSwingOOP
View code on GitHub ↗
Research · Chengdu, 2024

Sustainable Energy Thesis

Modeled the efficiency, feasibility and policy impact of integrating renewable energy in urban China, then presented the findings to an expert panel.

ResearchData-driven analysisPresentation
Leadership · 2021 – Present

Action for Mothers and Children

$68K+ raised for women's and children's healthcare in Kosovo

I coordinate fundraising events with nonprofit organizers and lead donor outreach across the tri-state area. The work has helped establish multiple Women's Health Resource Centers.

04 / skills

Skills

languages
PythonJavaCC++SQLARM64 Assembly
ai & data
Databricks (Genie)Microsoft CopilotClaude CodeAI agent developmentPrompt engineeringData analysis & visualization
tools
Git & GitHubLinux / BashDigital logic & circuitsSalesforceMicrosoft Office & SharePoint
spoken
EnglishAlbanianSpanish (basic)
05 / education

Stevens Institute of Technology

Degree
B.S. Computer Science
Minor
Machine Learning
GPA · Graduating
3.8 · May 2029
relevant coursework
CS 385 AlgorithmsCS 284 Data StructuresCS 382 Computer ArchitectureCS 135 Discrete StructuresLinear AlgebraProbability & StatisticsMultivariable CalculusFrontiers of Technology: AI & MLCS 101 Research & Entrepreneurship

Member of the Stevens Artificial Intelligence Club since Sep 2025.

06 / contact

Let's talk.

I'm open to internships and co-ops in AI, data science and software engineering. Email is the best way to reach me.