AI Engineer with 1+ years in Machine Learning, NLP, and Data Science.
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Mudith Pathirana is an aspiring AI Developer with 1.0 years of contract experience in automating forex market analysis and developing algorithmic trading bots. Currently pursuing a BSc (Hons) in Artificial Intelligence and Data Science, Mudith has strong skills in Python, Machine Learning, NLP, MLOps, and cloud platforms like AWS. Their project portfolio includes real-time trade copying systems, data-driven platforms, and cloud-native ETL pipelines, demonstrating a practical approach to AI and data engineering challenges.
Informatics Institute of Technology (IIT)
BSc (Hons) · Artificial Intelligence and Data Science
August 1, 2023 – June 30, 2028
Royal College Polonnaruwa
G.C.E Advanced Level · Commerce Stream
June 1, 2017 – May 31, 2022
Royal College Polonnaruwa
Ordinary Level
June 1, 2017 – May 31, 2022
Global Links Trading
AI Developer (Contract)
January 1, 2024 – January 1, 2025
UAE
Skill-Sync Data-Driven Platform
June 24, 2026 – Present
Manually extracted data from Sri Lankan universities and built a dataset of 1,000 rows of course curricula per degree programme, including curated skills learned through each course. Fine-tuned a pre-trained TSDAE (Transformer-based Sequential Denoising Auto-Encoder) model to calculate cosine similarity; performed text preprocessing including lemmatization for Jaccard similarity analysis. Developed curriculum relevance metrics comparing academic skills with market demands using Cosine and Jaccard similarity measures to determine alignment scores per university. Integrated alignment scores via a Flask API to deliver results to the front-end interface.
Cloud-Native AI Job Market Data Pipeline
June 24, 2026 – Present
Designed and implemented a robust cloud-native ETL pipeline orchestrating automated ingestion of global job market datasets into an Amazon S3 data lake using Apache Airflow for workflow management. Developed serverless data processing and feature engineering workflows with AWS Glue and PySpark to transform raw unstructured data into refined metrics including salary bands and technical skill counts. Established a high-performance analytical feature store using Amazon RDS PostgreSQL to ensure data integrity and high availability for downstream ML and data visualisation requirements.
Algo Trading Pipeline
June 24, 2026 – Present
Developed an automated algorithmic trading bot integrating MetaTrader 5 for live execution and Backtrader for performance testing against Dukascopy historical tick data. Implemented a multi-timeframe three-candle pattern recognition strategy (M5, H1, D1) with dynamic, zone-based risk management. Applied a fixed 1.5:1 risk-to-reward ratio framework to enforce consistent capital allocation and disciplined trade sizing across all executed positions.
AI Tutor – RAG Web Application
June 24, 2026 – Present
Developed a Retrieval-Augmented Generation (RAG) web application using Streamlit, LangChain, and Google Gemini 1.5 Flash to transform academic documents into structured study materials. Integrated ChromaDB and HuggingFace embeddings (all-MiniLM-L6-v2) for source-grounded semantic search and context retrieval. Designed a custom document-parsing pipeline supporting PDF, DOCX, and TXT formats alongside an FPDF2-powered markdown-to-PDF export engine.
Text Spam Classification
June 24, 2026 – Present
Built a a multi-input text classification pipeline for spam detection, measured by ROC-AUC, precision, recall, and accuracy evaluated on a 20% holdout test set. Performed exploratory data analysis and engineered metadata features including message length and character-level counts to supplement raw text signals for improved model discrimination. Built a custom Keras text vectorization layer and trained a TensorFlow neural network with balanced class weights to robustly address class imbalance throughout the classification pipeline.
Trade Copier
June 24, 2026 – Present
Developed a real-time trade copying system using Python and MetaTrader 5 (MT5) that automatically detects new positions and broadcasts trade details including volume, stop-loss, and take-profit parameters via a PubNub messaging infrastructure. Engineered the consumer-side architecture to subscribe to trade feeds, dynamically parse JSON payloads, and execute mirrored positions instantly via the MT5 API. Implemented custom order routing and execution logic to ensure accurate and low-latency trade replication across all subscriber accounts.
Predictive Customer Churn Analysis
June 24, 2026 – Present
Developed a predictive modeling solution for a telecom firm by performing end-to-end data cleaning, feature engineering, and preprocessing on a dataset of over 7,000 records. Benchmarked Decision Tree and Neural Network architectures to identify the most effective classification model for customer retention prediction. Validated predictive accuracy using ROC-AUC metrics and classification reports to ensure reliable and ethical data-driven insights.
IMDB Movie Sentiment Classification
June 24, 2026 – Present
Build a binary sentiment classification on the IMDB Movie Reviews dataset, achieving 85% validation accuracy, by implementing a Gated Recurrent Unit (GRU) architecture in PyTorch. Engineered a custom tokenization and sequence-padding pipeline using torchtext to preprocess raw review text into fixed-length numerical sequences suitable for sequential RNN input. Designed an end-to-end deep learning pipeline incorporating embedding layers, GRU cells, and binary cross-entropy optimization to enable context-aware, word-by-word text understanding.
Cultural Fit Analysis
The candidate demonstrates a strong interest in AI and data science, aligning well with an AI Engineer role. The variety of projects, from financial trading bots to educational platforms and market analysis, indicates a broad curiosity and willingness to tackle diverse problem domains. The use of various technologies (AWS, Google Gemini, Hugging Face, PyTorch, TensorFlow) suggests an openness to new tools and continuous learning. The remote contract experience also points to self-discipline and autonomy, which are valuable for cultural fit in many modern tech environments. However, the candidate's experience level (1 year) suggests they are still early in their career, and their cultural fit would benefit from further assessment during interviews.
Soft Skills & Operational Fit
The candidate's resume highlights soft skills such as Project Management, Teamwork, and Critical Thinking. The diverse project portfolio, including collaborative efforts like the 'Skill-Sync Data-Driven Platform' and 'Cloud-Native AI Job Market Data Pipeline', suggests an ability to work in team settings and manage project lifecycles. The contract role as an AI Developer also indicates adaptability and a results-oriented approach. However, without specific behavioral assessment data, the depth of these soft skills and operational fit cannot be fully ascertained.