{"name":"Elyesa Tee Way Yien","lastVerified":"2026-08-07","role":"Software & Data Engineer","location":"Puchong, Selangor, Malaysia","email":"etee3001@gmail.com","phone":"+6012-921 3001","links":[{"label":"LinkedIn","href":"https://www.linkedin.com/in/elyesa-tee-865536320"},{"label":"GitHub","href":"https://github.com/veyroxie"}],"siteHref":"https://veyroxie.github.io/Personal-Website/","sourceHref":"https://github.com/veyroxie/Personal-Website","experience":[{"company":"CobiNeural","title":"Digital Projects Intern to Engineer (Part-Time)","meta":"software & data · offered full-time conversion · Dec 2025 – Present","claims":[{"text":"Built the agentic backend for the company's AI energy-analytics chatbot: a LangGraph pipeline with a fetch-then-analyze flow on the Claude SDK and Model Context Protocol tools over FastAPI, so it handled messy real-world queries instead of breaking on anything not explicitly coded for.","evidence":{"tier":"attested","reason":"Production codebase at CobiNeural, confidential","stack":["LangGraph","Claude SDK","MCP","FastAPI"],"metrics":[]}},{"text":"Shipped an anti-hallucination suite to production: provenance tracking (EvidencePack/StepResult), a numerical-claim validator that flagged ungrounded numbers, a 5-tier reliability scale, and a citations parser/resolver, so analytics answers could be traced back to the source sensor data.","evidence":{"tier":"attested","reason":"Production codebase at CobiNeural, confidential","stack":["provenance tracking","claim validation"],"metrics":["28 tests"],"history":[{"date":"2026-08-07","note":"employer name corrected from pseudonym to CobiNeural"}]}},{"text":"Built a staged extract/transform/load toolkit for correcting production ClickHouse sensor data: every edit happens in parquet, a single load module is the only ClickHouse write surface, and delete-plus-reaggregation is emitted as reviewable SQL the toolkit never executes. Ingest streams one sensor at a time behind a server-memory gate, after an earlier batch-loading approach overloaded the server.","evidence":{"tier":"artifact","stack":["ClickHouse","Parquet","Athena","S3","Python"],"metrics":["6.7B rows","2,288 sensors","50 buildings","16 workflow wrappers","8 rollup tables"],"history":[{"date":"2026-08-07","note":"attested → artifact: scale figures verified against system.parts and a 30-day active-sensor query"},{"date":"2026-08-05","note":"PR #318 merged to main after review (106 files)"},{"date":"2026-07-24","note":"claim added, scoped to dedup and threshold filtering only"}]}},{"text":"Solved correction-versus-dedup correctness on a ReplacingMergeTree table by stamping ingested_at at insert so a correction deterministically wins FINAL, with time-spread read-back verification gating every run before it reports success.","evidence":{"tier":"attested","reason":"Production codebase at CobiNeural, confidential","stack":["ClickHouse","ReplacingMergeTree"],"metrics":[],"history":[{"date":"2026-08-07","note":"claim added, split out of the ETL toolkit claim"}]}},{"text":"Prototyped and tested a statistical anomaly detector (GESD, MAD/robust-z, IQR fences, per-sensor delta floor, spike-cap safeguard) covering degenerate cases, kept as an option for active threshold filtering, alongside confirmation-phrase and pre-delete backup guards on destructive data paths.","evidence":{"tier":"attested","reason":"Production codebase at CobiNeural, confidential","stack":["GESD","MAD / robust-z","IQR"],"metrics":["16 unit tests"]}},{"text":"Authored the EECA compliance-report engine (chat-driven editing, browser-editable output, soft delete with auto-expiry) and built its frontend in SvelteKit and Svelte 5: streaming responses, inline artifact cards, and ECharts visualizations.","evidence":{"tier":"attested","reason":"Production codebase at CobiNeural, confidential","stack":["SvelteKit","Svelte 5","ECharts"],"metrics":["~2,000 lines"]}}]}],"projects":[{"name":"StudyHub","tagline":"Tuition-Centre Management Platform · Go · vanilla JS · PostgreSQL","description":"Solo-built and deployed to production: role-based dashboards, billing/payroll, attendance, analytics, notifications, and JWT auth with email verification.","evidence":{"tier":"live","href":"https://studyhub.fit","linkLabel":"studyhub.fit","stack":["Go (chi, pgx)","PostgreSQL","WebSocket","JWT"],"metrics":["143 REST routes","25-table schema","81 backend tests","~22k LOC"]}},{"name":"ollama-chatbot","tagline":"Agentic learning sandbox · Python","description":"A personal sandbox for agent patterns: a ReAct loop with tool execution, retries, and streaming across both Ollama and Gemini. Where I prototype the ideas the CobiNeural work productionises.","evidence":{"tier":"live","href":"https://github.com/veyroxie/ollama-chatbot","linkLabel":"source on GitHub","stack":["Python","ReAct","tool execution"],"metrics":[]}},{"name":"Crime in Malaysia","tagline":"Interactive Data Stories · Vega-Lite · JavaScript","description":"Two single-page visual narratives built from raw public CSVs: a choropleth world map, parallel-coordinates crime trends, a state homicide stream graph, radial prisoner-index charts, and case heatmaps.","evidence":{"tier":"live","href":"https://veyroxie.github.io/dv2/","linkLabel":"live visualisation","stack":["Vega-Lite","JavaScript"],"metrics":[]}},{"name":"Multiplayer Horror Game","tagline":"Co-op terminal game · Java · OOP","description":"A terminal-controlled co-op game (Lethal Company clone) built with three teammates. Wrote the enemy AI behaviour and loot mechanics.","evidence":{"tier":"live","href":"https://github.com/veyroxie/lethal-company-knockoff","linkLabel":"source on GitHub","stack":["Java","OOP"],"metrics":[]}},{"name":"Deep Learning for NLP","tagline":"Sequence models · PyTorch","description":"Trained RNN, LSTM, and Transformer models with attention for text generation and sequence classification, then compared how each performed.","evidence":{"tier":"artifact","stack":["PyTorch","RNN","LSTM","Transformer"],"metrics":[]}}],"skills":[{"label":"Languages","items":["Python","Go","TypeScript","JavaScript","Java","SQL","HTML/CSS"]},{"label":"AI / ML","items":["Claude SDK","LangGraph","Model Context Protocol","PyTorch","Anomaly Detection (GESD/MAD)"]},{"label":"Backend","items":["Go (chi, pgx)","FastAPI","PostgreSQL/pgvector","ClickHouse","GraphQL","MongoDB","WebSocket","JWT"]},{"label":"Frontend","items":["SvelteKit","Svelte 5","Tailwind","ECharts","Vega-Lite","Tableau"]}],"education":[{"title":"Monash University Malaysia","detail":"B.Sc. Computer Science (Data Science) · 2022 – Present"},{"title":"INTI International College Subang","detail":"Cambridge A Levels · Merit Scholarship · 2020 – 2021"},{"title":"Nobel International School","detail":"Cambridge IGCSE · ICE (Distinction) · 2019"}],"leadership":[{"title":"Monash E-Sports Club","detail":"Secretary (current) and former Social Media Manager. Sorted out leadership friction, kept official paperwork on track, and ran promo content and inter-club relationships."}]}