NLP & AI/ML Engineer

Jarsham Dhir

Specializing in

Montreal, QC, Canada

Cross-Lingual NLP

Bilingual word vector alignment across English & Punjabi using FastText & MUSE

LLM Evaluation

Production-scale AI annotation & evaluation at TELUS Digital for LLM training pipelines

ML Mathematics

3× certified by Imperial College London — Linear Algebra, Calculus & PCA

Featured Work

NLP research and ML engineering projects

Monolingual Word Embeddings

Low-resource Punjabi NLP model training

  • Trained Word2Vec and FastText models on Punjabi text corpora for semantic and morphological word representation
  • Tuned hyperparameters — window size, vector dimensions, epochs — to optimize word capture for a low-resource language
PythonWord2VecFastTextGensimNLP
View on GitHub →

Comparable Sentence Mining

Parallel corpus extraction for English–Punjabi

  • Built end-to-end pipeline to extract semantically comparable English–Punjabi sentence pairs using cross-lingual embeddings and cosine similarity
  • Applied Unicode normalization and custom Gurmukhi script tokenizers for robust multilingual text processing
Pythonscikit-learnCosine SimilarityNLPPipelines
View on GitHub →

Technical Skills

ML & AI Frameworks

TensorFlowPyTorchscikit-learn FastTextWord2VecGensim

NLP Domains

Word EmbeddingsCross-Lingual AI LLM EvaluationData Annotation Machine TranslationText Mining

Languages & Tools

PythonSQLJavaScript DockerGitGitHub

Data & Analysis

PandasNumPyJupyter Data CleaningPipeline Dev

Experience

Online Data Analyst

TELUS Digital · Freelance
Jul 2025 – Jan 2026
  • Evaluated and annotated AI-generated outputs (text, search results, conversations) to support LLM training and fine-tuning at production scale
  • Applied structured annotation guidelines across multilingual datasets, assessing accuracy, relevance, fluency, and safety of model responses
  • Delivered high-quality feedback directly informing model improvement cycles for one of Canada’s largest telecom-backed AI platforms
  • Maintained annotation accuracy and throughput targets over a 7-month remote engagement

Data Analyst Intern

HowToAbroad · Remote
May 2022 – Jun 2022
  • Analyzed user engagement datasets using Python (Pandas, NumPy) and SQL to surface actionable insights for content and product decisions
  • Built data cleaning pipelines to normalize and transform audience metrics, improving reporting consistency
  • Produced weekly performance dashboards informing marketing strategy for an internationally-focused immigration content platform

Education & Certifications

Education

Diploma in Network Management

LaSalle College, Montreal, QC

2023 – 2025

Network Infrastructure · Cloud Computing · Cybersecurity · IT Systems

Bachelor of Computer Applications (BCA)

Guru Nanak Dev University, Amritsar, India

2018 – 2021

Software Development · Data Structures · Algorithms · Databases · Networks

Certifications

ICL

Mathematics for Machine Learning: PCA

Imperial College London · Coursera · Aug 2024

Credential ID: 7MTSOAFT1VKG

ICL

Mathematics for Machine Learning: Multivariate Calculus

Imperial College London · Coursera · Mar 2022

Credential ID: JC3VY6KET7BM

ICL

Mathematics for Machine Learning: Linear Algebra

Imperial College London · Coursera · Apr 2022

Credential ID: 69R5MGWJAH3F

About Me

Entry-level Software Developer and AI/ML Engineer with hands-on experience in NLP, multilingual word embeddings, and AI data annotation for production systems. Skilled in Python, TensorFlow, PyTorch, and Docker.

I work at the intersection of language and machine learning — building systems that understand text across linguistic boundaries. My focus on low-resource Punjabi NLP is both a technical challenge and a personal mission to make AI more accessible across languages.

🇬🇧
English Fluent
🇮🇳
Punjabi Native
🇮🇳
Hindi Conversational

Open to full-time, contract, and co-op roles in AI/ML, NLP, and Software Engineering across Canada.