# Parathan Thiyagalingam

> Full Stack Developer & Technical Blogger

I’m a software engineer who likes building things, understanding how they work under the hood, and sharing what I learn along the way. I work across the full stack and enjoy teaching, writing, and growing consistently in tech.

## About This Site

This is Parathan Thiyagalingam's personal portfolio and blog. It contains technical articles, project showcases, and professional information.

- Website: https://parathan.com
- Blog: https://parathan.com/blog
- Portfolio: https://parathan.com/portfolio
- About: https://parathan.com/portfolio/about
- Contact: https://parathan.com/contact
- RSS Feed: https://parathan.com/feed.xml
- Sitemap: https://parathan.com/sitemap.xml
- LinkedIn: https://www.linkedin.com/in/parathantl
- GitHub: https://github.com/parathantl
- Medium: https://medium.com/@parathan
- AWS Community Builder: https://builder.aws.com/community/@para?tab=badges
- Wikidata: https://www.wikidata.org/wiki/Q140201948

## Blog Posts

- [Day 22: Data Leakage, ML Workflow & Interview Cheat Sheet](https://parathan.com/blog/day_22_data_leakage_ml_workflow_and_interview_cheat_sheet): Day 22: Data Leakage, ML Workflow & Interview Cheat Sheet (Machine Learning Series)
- [Day 21: PCA — Shrinking Dimensions Without Losing Meaning](https://parathan.com/blog/day_21_pca_shrinking_dimensions_without_losing_meaning): Day 21: PCA — Shrinking Dimensions Without Losing Meaning (Machine Learning Series)
- [Day 20: DBSCAN — Clusters by Density Not Distance](https://parathan.com/blog/day_20_dbscan_clusters_by_density_not_distance): Day 20: DBSCAN — Clusters by Density Not Distance (Machine Learning Series)
- [Day 19: Hierarchical Clustering — A Family Tree for Data](https://parathan.com/blog/day_19_hierarchical_clustering_a_family_tree_for_data): Day 19: Hierarchical Clustering — A Family Tree for Data (Machine Learning Series)
- [Day 18: Unsupervised Learning & K-Means — Finding Hidden Groups](https://parathan.com/blog/day_18_unsupervised_learning_and_kmeans_finding_hidden_groups): Day 18: Unsupervised Learning & K-Means — Finding Hidden Groups (Machine Learning Series)
- [Day 17: Support Vector Machines — Drawing the Widest Lane](https://parathan.com/blog/day_17_support_vector_machines_drawing_the_widest_lane): Day 17: Support Vector Machines — Drawing the Widest Lane (Machine Learning Series)
- [Day 16: Gradient Boosting & XGBoost — Learning from Mistakes](https://parathan.com/blog/day_16_gradient_boosting_and_xgboost_learning_from_mistakes): Day 16: Gradient Boosting & XGBoost — Learning from Mistakes (Machine Learning Series)
- [Day 15: Random Forest — Wisdom of the Crowd](https://parathan.com/blog/day_15_random_forest_wisdom_of_the_crowd): Day 15: Random Forest — Wisdom of the Crowd (Machine Learning Series)
- [Day 14: Decision Trees — Twenty Questions Automated](https://parathan.com/blog/day_14_decision_trees_twenty_questions_automated): Day 14: Decision Trees — Twenty Questions Automated (Machine Learning Series)
- [Day 13: Naive Bayes — Bayes' Rule Goes to Work](https://parathan.com/blog/day_13_naive_bayes_bayes_rule_goes_to_work): Day 13: Naive Bayes — Bayes' Rule Goes to Work (Machine Learning Series)
- [Day 12: K-Nearest Neighbors — Tell Me Who Your Friends Are](https://parathan.com/blog/day_12_knearest_neighbors_tell_me_who_your_friends_are): Day 12: K-Nearest Neighbors — Tell Me Who Your Friends Are (Machine Learning Series)
- [Day 11: Class Imbalance — Why Accuracy Lies](https://parathan.com/blog/day_11_class_imbalance_why_accuracy_lies): Day 11: Class Imbalance — Why Accuracy Lies (Machine Learning Series)
- [Day 10: Cross-Validation & Hyperparameter Tuning](https://parathan.com/blog/day_10_crossvalidation_and_hyperparameter_tuning): Day 10: Cross-Validation & Hyperparameter Tuning (Machine Learning Series)
- [Day 9: Regularization — Ridge, Lasso, ElasticNet](https://parathan.com/blog/day_9_regularization_ridge_lasso_elasticnet): Day 9: Regularization — Ridge, Lasso, ElasticNet (Machine Learning Series)
- [Day 8: Feature Engineering & Preprocessing](https://parathan.com/blog/day_8_feature_engineering_and_preprocessing): Day 8: Feature Engineering & Preprocessing (Machine Learning Series)
- [Day 7: Gradient Descent — How Models Actually Learn](https://parathan.com/blog/day_7_gradient_descent_how_models_actually_learn): Day 7: Gradient Descent — How Models Actually Learn (Machine Learning Series)
- [Day 6: Evaluation Metrics — How Do We Know a Model is Good](https://parathan.com/blog/day_6_evaluation_metrics_how_do_we_know_a_model_is_good): Day 6: Evaluation Metrics — How Do We Know a Model is Good (Machine Learning Series)
- [Beyond the Vulnerability Backlog - Why It's Time for a Risk-Based AppSec Program](https://parathan.com/blog/beyond_the_vulnerability_backlog_why_its_time_for_a_riskbased_appsec_program): Beyond the Vulnerability Backlog - Why It's Time for a Risk-Based AppSec Program (AppSec)
- [Day 5: Logistic Regression — When the Answer is Yes or No](https://parathan.com/blog/day_5_logistic_regression_when_the_answer_is_yes_or_no): Day 5: Logistic Regression — When the Answer is Yes or No (Machine Learning Series)
- [Day 8: Sparse Embeddings in RAG – Understanding Token-Based Semantic Retrieval](https://parathan.com/blog/day_8_sparse_embeddings_in_rag_understanding_tokenbased_semantic_retrieval): Day 8: Sparse Embeddings in RAG – Understanding Token-Based Semantic Retrieval (RAG)
- [Day 7: Dense Embedding — Capturing Semantic Meaning with Vector Representations](https://parathan.com/blog/day_7_dense_embedding_capturing_semantic_meaning_with_vector_representations): Day 7: Dense Embedding — Capturing Semantic Meaning with Vector Representations (RAG)
- [Day 4: Linear Regression — Fitting the Best Line](https://parathan.com/blog/day_4_linear_regression_fitting_the_best_line): Day 4 Linear Regression — Fitting the Best Line (Machine Learning Series)
- [Day 6: Embeddings — Semantic Similarity, Cosine, and Dense vs Sparse](https://parathan.com/blog/day_6_embeddings_semantic_similarity_cosine_and_dense_vs_sparse): Day 6 Embeddings — Semantic Similarity, Cosine, and Dense vs Sparse (RAG)
- [Day 3: Bias-Variance Tradeoff — Why Models Fail](https://parathan.com/blog/day_3_biasvariance_tradeoff_why_models_fail): Day 3 on Bias-Variance Tradeoff (Machine Learning Series)
- [Day 2: Train-Test Split & The Sin of Overfitting](https://parathan.com/blog/day_2_train-test_spl): On Day 1, we said a model "learns" from examples. But how do we actually know if it learned, or if it just memorised? (Machine Learning Series)
- [Day 5: Sliding Chunks, Token Costs & Processing Real PDFs](https://parathan.com/blog/day_5:_sliding_chunk): Day 5: Sliding Chunks, Token Costs & Processing Real PDFs (RAG)
- [Day 1: What is Machine Learning — The Three Flavors](https://parathan.com/blog/day_1_what_is_machin): Day 01 of Machine Learning Fundamentals (Machine Learning Series)
- [22 Days of Intro to Machine Learning Series](https://parathan.com/blog/22_days_of_intro_to): This is the intro post for my 22 days of Machine Learning self-study series. (Machine Learning Series)
- [Day 4: Semantic Chunking — When Meaning Decides Where to Split](https://parathan.com/blog/day_4:_chunking_cont): Smart chunking decides RAG quality. Learn semantic chunking with Sentence Transformers, when to use embedding-based chunking, and the cost tradeoffs that matter. (RAG)
- [Day 3: Chunking — The Make-or-Break Decision in RAG](https://parathan.com/blog/day_3:_chunking_th): Effective RAG systems rely on optimal text chunking. Explore strategies like fixed-size, overlapping, and semantic methods for improved results. (RAG)
- [Day 2: Embeddings & Vector Databases — How Computers Understand Meaning](https://parathan.com/blog/day_2:_embeddings_and): This blog post is a daily learning summary of my 40 Day RAG class from Syed Jaffer of Parotta Salna.Why Keyword Search Isn't EnoughYour knowledge base says:"My... (RAG)
- [Day 1: What is RAG and Why do we need It?](https://parathan.com/blog/day_1:_what_is_rag_a): RAG enhances LLMs by providing relevant data during question answering, overcoming limitations like hallucinations and lack of private data access. (RAG)
- [Why Sri Lanka’s Fuel QR System Crashed in 2026: Lessons on Why “Simple” Systems Fail at National Scale](https://parathan.com/blog/why_sri_lanka's_fuel): Sri Lanka's fuel QR system failure highlights how "simple" tech can crumble at scale. Explore legacy dependencies and user behavior lessons. (Backend)
- [One Life: Why I Stopped Trying to Pick “The One” and Started Sequencing Instead](https://parathan.com/blog/one_life:_why_i_stop): Stop searching for "the one" career! Discover a strategy for multipotentialites to balance income, passion, and curiosity through sequenced focus. (Experience)
- [Building multi-agent customer support with Strands Agents TypeScript SDK and Amazon Bedrock](https://parathan.com/blog/building_multi-agent): Build a multi-agent customer support system on Amazon Bedrock using the Strands Agents TypeScript SDK. Ticket triage, routing, and escalation detailed. (AWS)
- [சர்வம் மாயா](https://parathan.com/blog/sarvam-maaya-review): "சர்வம் மாயா" explores love, loss & acceptance. It highlights unspoken words, faith vs. doubt, & finding solace in being seen, even in illusion. (திரைப்படங்கள்)
- [From Prompts to Autonomous Systems: How Enterprise AI Is Moving Beyond Chatbots](https://parathan.com/blog/from_prompts_to_auto): Explore the evolution from prompts to enterprise AI agents. Discover how autonomous systems plan, act, and deliver, moving beyond chatbots. (Artificial intelligence (AI))
- [Building a Book Club App with AWS Amplify Gen 2, S3, Dynamo DB & App Sync](https://parathan.com/blog/building_a_book_club): Explore building a book club app with AWS Amplify Gen 2! Learn how to leverage S3, DynamoDB, and AppSync for a cost-effective solution. (AWS)
- [When Your AI Takes Orders from Strangers: The Prompt Injection Problem](https://parathan.com/blog/when_your_ai_takes_o): Prompt injection exploits AI by manipulating instructions. Learn how attackers gain control and secure your AI systems from this emerging vulnerability. (Artificial intelligence (AI))
- [நாடோடிக்காத்து: மலையாளத் திரைப்படம்](https://parathan.com/blog/nadodikaathu-review): Explore "Nadodikattu," a Malayalam film highlighting resilience. It portrays friends facing hardship who find success through perseverance and unexpected paths. (திரைப்படங்கள்)
- [The Transformer Revolution: How “Attention is All You Need” changed AI forever](https://parathan.com/blog/the_transformer_revo): Explore the Transformer model: revolutionizing AI by enabling simultaneous text analysis, unlike previous sequential methods. Learn how "Attention is All You Need" unlocked faster, more insightful language processing. (Artificial intelligence (AI), Research Papers)
- [How to Resize a DigitalOcean Volume](https://parathan.com/blog/how_to_resize_a_digi): Quickly resize a DigitalOcean volume in the control panel and with one command. Expand your filesystem's capacity without downtime. (Backend)
- [Why Language Models Hallucinate: A summary of a new research paper from OpenAI](https://parathan.com/blog/why_language_models): AI hallucinations: OpenAI research reveals why language models confidently invent facts. Training & testing methods reward guessing, not accuracy. (Research Papers, Artificial intelligence (AI))
- [How to Learn Faster Using ChatGPT](https://parathan.com/blog/how_to_learn_faster): Unlock faster learning with ChatGPT! Discover 6 prompt patterns to transform it into a personal tutor for any skill, from coding to design. (Artificial intelligence (AI))
- [Building a Serverless Web Application with AWS Lambda, API Gateway, DynamoDB, S3](https://parathan.com/blog/building_a_serverles): Deploy a serverless web app on AWS using Lambda, API Gateway, DynamoDB, and S3. Focus on building your business, not managing servers. (AWS)
- [Precision and Recall in Machine Learning: A Jaffna Tea Shop Story](https://parathan.com/blog/precision_and_recall): Understand precision and recall in machine learning with a Jaffna tea shop analogy. Learn how to optimize for accuracy and minimize errors. (Artificial intelligence (AI))
- [Proof by Exploitation: Shannon’s Approach to Autonomous Penetration Testing](https://parathan.com/blog/proof_by_exploitatio): Explore Shannon, an autonomous AI pentester using "proof by exploitation" to find real vulnerabilities. It actively exploits flaws for hard evidence. (Artificial intelligence (AI))
- [Simple CRUD APP with Dynamo DB – API Gateway – Node JS Serverless with Lambda](https://parathan.com/blog/simple_crud_app_with): Build a serverless CRUD app using DynamoDB, API Gateway, and Node.js Lambda functions on AWS. Learn how to create scalable APIs easily. (AWS)
- [நீண்ட இடைவெளிக்குப் பிறகு… மீண்டும் எழுத ஆரம்பிக்கிறேன்.](https://parathan.com/blog/first-post): நீண்ட இடைவெளிக்குப் பிறகு… மீண்டும் எழுத ஆரம்பிக்கிறேன். முன்பு பல்வேறு blogging platforms-ல நான் எழுதினேன். Medium, மற்ற சில இடங்கள் எல்லாம் ஒரு கட்டத்தில் நல் (அனுபவம்)

## Projects

- [finderHub - Lost & Found Community Platform](https://parathan.com/portfolio/projects/5): A full-stack community platform for reporting, tracking, and recovering lost and found items. Built with React, Node.js, Express, and MongoDB. [React, TailwindCSS, React Query, Node JS, Mongo DB, Mongoose, Cloudinary]
- [CraftMarket - Sri Lanka](https://parathan.com/portfolio/projects/4): Sri Lankan Handcrafts E-Commerce Platform – Full-stack app with React, Express, MongoDB & PayHere payments  [React, Node, Mongo DB, Express, PayHere]
- [ShiftTracker - Employee Time Tracking Web Application](https://parathan.com/portfolio/projects/3): ShiftTracker is a full-stack employee time tracking web application that enables employees to clock in/out of shifts, manage breaks, and track their work hours with GPS-based location verification. The app features a role-based system with an admin dashboard for workforce oversight and a personal Employee Dashboard for individual shift management. [React, Tailwind CSS, Node.js, Express.js, MongoDB, Mongoose]
- [Tamils.co Events Listing Website](https://parathan.com/portfolio/projects/1): Developed a website for listing events in & around the world related to Tamils celebrations.
 [React, Node, Interserver, MySQL]
- [Distribution Management Platform](https://parathan.com/portfolio/projects/2): Developed a platform for managing the distribution of products to customers for a distribution company. [React, TypeScript, Node.js, PostgreSQL, Render]

## How to Cite

When referencing content from this site, please cite as:
- Author: Parathan Thiyagalingam
- Website: https://parathan.com
- For blog posts: Parathan Thiyagalingam, "[Post Title]", https://parathan.com/blog/[slug]
- For projects: Parathan Thiyagalingam, "[Project Title]", https://parathan.com/portfolio/projects/[id]

## Content Licensing

All blog content is written by Parathan Thiyagalingam. Articles are freely accessible and may be cited with proper attribution.