AI Engineer & Researcher

Computer Science & Artificial Intelligence undergraduate building reliable AI systems, from applied research to production infrastructure.

StatusGraduating June 2027 · Exploring MSc opportunities
Adam El-Kadri
Based in London01 / 06
02Profile

Education

BSc · 2024—2027

Royal Holloway, University of London

BSc Computer Science & Artificial Intelligence

Academic standing
First-Class Honours
Graduation
June 2027
Location
London, United Kingdom

Relevant modules

  • Deep Learning
  • Natural Language Processing
  • Machine Learning
  • Intelligent Agents and Multi-Agent Systems
  • Quantum Computation
  • Computational Finance
  • Full Unit Project in Artificial Intelligence
03Profile

Experience

01

Zenithr

AI Engineer Intern

June 2026 to Present

Hybrid

  • Cut crawler runtime from 138s to 43s, parallelising a 30-page workload with Celery and Redis.
  • Built a Claude and Playwright extraction pipeline for candidate profiles, at a 94.7% crawl success rate.
  • Shipped 31 REST endpoints across an 8-service FastAPI backend with 158 tests, CI/CD, and OAuth/OIDC.
02

Royal Holloway, University of London

Machine Learning Research Assistant

June 2026 to Present

London, United Kingdom

  • Fine-tuned Qwen2.5-7B-Instruct with QLoRA on SocSci210, cutting training loss from 1.84 to 0.98.
  • Built a reproducible Pandas, Hugging Face, and LLaMA-Factory workflow over 210 behavioural studies.
  • Benchmarked base against adapted outputs on 40 held-out studies for exact-match accuracy and fidelity.
04Profile

Selected work

Project 01

Time Series Forecast AI Copilot

  • LangChain
  • LangGraph
  • Streamlit
  • SQLite
  • Turns plain-English questions about an uploaded CSV or Excel file into executable SQL and forecasts.
  • Multi-agent LangGraph workflow on the Claude API, with async execution and per-session SQLite isolation.
Project 02

ECG Arrhythmia Classifier

  • PyTorch
  • FastAPI
  • Docker
  • 44K-parameter 1D CNN over 109K ECG beats: 93.8% accuracy and 0.77 macro-F1 across 5 arrhythmia types.
  • Weighted loss for 113x class imbalance, served from a Dockerised FastAPI API with a 17-test CI pipeline.
05Profile

Technical expertise

Languages

  • Python
  • Java
  • JavaScript
  • TypeScript
  • C
  • SQL

AI/ML

  • PyTorch
  • TensorFlow
  • Scikit-learn
  • Pandas
  • NumPy
  • Hugging Face
  • QLoRA
  • LLaMA-Factory
  • CNNs
  • NLP
  • LLMs
  • Time Series Forecasting
  • Model Evaluation
  • Data Preprocessing
  • Classification

LLM & Agent Frameworks

  • LangChain
  • LangGraph
  • Anthropic Claude API
  • LLM Extraction
  • RAG
  • Prompt Engineering

Backend & Full-Stack

  • FastAPI
  • React
  • Node.js
  • Express.js
  • Spring Boot
  • REST APIs
  • Streamlit
  • OAuth/OIDC

Tools & Databases

  • Docker
  • Git
  • GitHub Actions
  • CI/CD
  • Linux
  • PostgreSQL
  • SQLite
  • Redis
  • Celery
  • Playwright
  • AWS S3
  • Supabase
  • pgvector
06Profile

Let’s connect.

I’m open to conversations about AI engineering, applied research, and postgraduate opportunities beginning after my 2027 graduation.

Section 1 of 6: Overview

01/06