Frontend development · Teaching · Applied AI

Hang Gu顾航

I teach computing at Jiangsu Vocational College of Finance and Economics. Previously a frontend lead, I build business interfaces, data tools and AI applications.

Explore my work ↓

Selected work

Three projects, with the implementation and its limits.

All 15 projects ↗

Crop inspection

FieldLens

Review an image, keep an inspection record, and export a report.

Region selection, comparison of records and optional AI assistance. The public demo works without signing in.

Try the demo ↗Case study GitHub

Source, scope & implementation

The public site uses preset demo data and does not collect passwords. Real sign-in needs the backend; inference needs an authorized model. Records stay in the browser.

Version 1.9 redesigns the homepage and working screens with readable type, four themes and semantic status colors. Detection, camera capture, review, reports and the compatible AI interface are retained. The twelve-case demo needs no account.

JavaScriptYOLOFlaskCanvasPlaywright

Preset data online; real inference needs a local model.

Counselor dashboard with synthetic Demo counselor account and tasks — Interface detail · Synthetic demo data

Education & collaboration

Counselor workspace

Organize tasks, messages and attachments around the people who handle them.

The public case explains role boundaries and release checks. Selected source files are available for review.

Read the case ↗Source excerpts Release checks

Source, scope & implementation

Public excerpts updated to v1.1.1 with frontend/backend paths and a refreshed 37-file source manifest. The case keeps its dated acceptance records; the excerpt is not a complete deployable system.

Vue 3Spring BootPlaywrightSHA-256

Local engineering case; public code is an excerpt.

Low-code development

Form studio

Keep the schema, form fields and validation rules in agreement.

Configure six field types and preview them immediately. Invalid imports leave the existing form intact.

Try the editor ↗Decisions & checks Source

Source, scope & implementation

Problem: Keep configuration, rendered fields and validation in agreement.
Public implementation: Six field types, stable keys, JSON import and an independent validation core.
Trade-off: Validate before replacing a schema; a valid import resets preview values.
Evidence: 10 core tests rerun on 1 Nov; inspect invalid-import recovery in the demo.

AI-assisted learning implementation with synthetic data. Client-side checks do not establish backend authorization or production acceptance.

SchemaDynamic formsValidationVue patterns

AI-assisted learning prototype with synthetic data.

New demo · MeteoFormer regional weather workspace. Explore a 24-hour synthetic forecast snapshot.

Try the weather demo ↗

More to explore: React workflows, data tools and a virtual presenter.

Browse demos & project films →

Research in progress

Two first-author submissions. Both are awaiting acceptance.

All research records ↗

Where do fusion gains come from?

Neurocomputing · submitted, not accepted

Reanalyzing archived predictions to distinguish fusion effects, operating points and dataset artifacts.

Read the study →

Climate risk in energy–water systems

Climate Risk Management · submitted, not accepted

Combining temporal modeling and regional connections to study multi-hazard risk and propagation.

Read the study →

Questions, contributions and public evidence →

Recent notes

Questions from reading and development, written down.

All 43 notes ↗

New study note2026.11.05

Energy meteorology: a reading path from data to decisions

A focused reading map for power forecasting, multi-source fusion and forecast correction, with a concrete question for each paper.

7 min readSources · questions · next steps

New study note2026.11.05

Before fusion: specify an energy-forecast baseline

Define the target, available inputs, chronological splits and error slices before comparing forecasting models.

8 min readSources · questions · next steps

New study note2026.11.05

A forecast dashboard should show what is known, and when

Apply frontend experience to forecast interfaces: distinguish observations from forecasts, stale responses from current ones, and unavailable data from zero.

7 min readSources · questions · next steps

Compiled from public code and project documents; examples, evidence and open questions are distinguished in each article.

Learning timeline

A learning timeline

Dates mark when notes were compiled; each entry links to the material.

11.052026
Reading & models
Energy meteorology · a starting point
A reading map, a baseline protocol and a dashboard design note.
11.042026
Engineering reflection
File inventory · define the boundary
A useful total starts with explicit scan rules and visible exclusions.
11.022026
Engineering reflection
Low-code · separate three states
Schema, user input and interface state should change for different reasons.
10.292026
Engineering reflection
AI and frontend · learn to review
Move beyond generating code: understand ownership, failure and verification.
10.272026
Reading & models
TFT · what is known at issue time?
Separate future-known inputs from measurements that arrive later.
10.202026
Engineering reflection
Async interfaces · handle late responses
Debouncing alone does not prevent an old result replacing a new one.

More learning repositories & earlier notes

NEW / REPRODUCIBILITY

From local model code to a readable study

Upstream attribution, early stopping, aligned predictions and evidence boundaries.

NEW / FINANCIAL PATTERNS

Ratios, missing values and cluster meaning

Undefined denominators, median imputation, standardization and PCA limits.

NEW / SPARK & QUALITY

From CSV audits to Spark transformations

Schemas, null semantics, deduplication, shuffle and event-time boundaries.

NEW / EVALUATION

Prediction ensembles and honest validation

Aligned sample IDs, metric averaging, thresholds, early stopping and test isolation.

01 / MODELS

From a task to an honest evaluation

Baselines, feature engineering, leakage, thresholds and error analysis.

02 / ANNOTATION

Label quality and disagreement

Coverage, paired agreement, guidelines and a conflict review queue.

03 / DATA COLLECTION

A bounded public-web crawler

Robots rules, same-origin scope, rate limits, JSONL and offline tests.

04 / DATA SYSTEMS

SQL, partitions and data quality

Window functions, shuffle, skew and reproducible transformation.

05 / FRONTEND

State and interface reliability

Form state, error recovery, keyboard access and performance.

06 / PRACTICE

Deployment decisions and evidence

Capacity tests, personal AI access and release checklists.

Experience

Work

2023.08 – Present

江苏财经职业技术学院

Lecturer & Teaching Section Head

Computing courses and practical work spanning data collection, storage, analysis and presentation.

2019.03 – 2022.07

Whale Cloud Technology

Frontend Lead

Configurable forms, workflow interfaces, reusable modules and team coordination.

2018.03 – 2019.02

Nanjing Siyue Network

Frontend Engineer

2015.12 – 2017.02

Wangdao Technology

Frontend Engineer

Education

2021 – 2023

Universiti Malaya

MSc · Computer Science, Data Science

2011 – 2015

Huaiyin Normal University

Bachelor's degree · Computer Science

Further applied work

Existing project experience also includes AI contract management and RAG retrieval tools (in progress), and a modular administrative dashboard (completed). Proprietary implementation details are not part of the public cases.

Technical practice & full profile

Technical practice

Research & modeling

Python · PyTorch · R
CLIP · Bayesian networks
Model evaluation & error analysis

Interface engineering

Vue · React · TypeScript
Forms · workflow states · API contracts
Accessibility · Playwright

Data & reproducibility

SQL · SQLite · Spark concepts
Quality checks & annotation
Versioned evidence & release gates

Get in touch

For project questions or research discussions, you can reach me by email.

Email Hang Gu ↗Read the CV ↗