Computer & Information Sciences
Data Science Challenges
Real data-science projects and challenges on Ewance — clean messy datasets, build and evaluate models, and turn raw data into decisions the way a working data scientist does. Solve them to build a portfolio of verified, recruiter-checkable proof you can do the work — not just describe it.
Recommended challenges
- CodeBeginnerNew
Implement a Constraint Solver for a Lisbon Tourism Scheduler
Model the next-week schedule as a CSP: variables are (guide, day, slot) assignments; domains are available routes; constraints encode language requirements, max consecutive tour…
- Constraint Satisfaction
- Backtracking Search
- Python
Introduction to Artificial Intelligence - CodeIntermediateNew
Train a Differentially Private Classifier on Medical Records
Use Opacus (PyTorch DP-SGD library). Train a tabular classifier (small MLP + gradient-boosted features) with DP-SGD at the agreed epsilon/delta. Run an accuracy-vs-privacy front…
- Differential Privacy
- Dp Sgd
- Opacus
Open coursework - CodeIntermediateNew
Detect Coordinated Fraud Rings via Link Analysis at a Neobank
You receive 90 days of account, login, and transaction data (around 1.2 million accounts, around 30 million events) plus a labeled set of 80 known fraud rings. Build a multi-rel…
- Graph Analysis
- Community Detection
- Link Analysis
Open coursework - CodeIntermediateNew
Defect Detection on PCBs for a Hardware-AI Manufacturer
Use the publicly-available PCB defect dataset (e.g., DeepPCB or HRIPCB). Fine-tune a small object detector (YOLOv8n or RT-DETR-small) on the 6 defect classes. Evaluate mean Aver…
- Object Detection
- Transfer Learning
- Model Evaluation
Computer Vision Practice your coursework on real scenarios.
Every challenge is shaped from real-world context — not generic exercises. The work mirrors what your degree prepares you for.
Why Ewance
- CodeFoundationalNew
Classify Retail Product Photos for an E-Commerce Marketplace
Use a publicly-available product-image dataset (e.g., Fashion-MNIST extended, or a Kaggle e-commerce subset of around 10k images across 12 categories). Fine-tune a small pretrai…
- Image Classification
- Transfer Learning
- Pytorch
Open coursework - ResearchIntermediateNew
Policy-Gradient Trading Agent on Historical Data
You receive 5 years of daily OHLCV (Open/High/Low/Close/Volume) data for 5 large-cap stocks. Build an episodic environment where each episode is one calendar year and the agent'…
- Policy Gradients
- Reinforce
- Rl Evaluation
Reinforcement Learning - ResearchIntermediateNew
Safety-Test a Customer-Service Agent for Adversarial Prompts
You receive a sandboxed instance of the agent (a tool-using LLM that can read account balances and open support tickets — both mocked). Design a red-team suite of at least 80 pr…
- LLM Agents
- Red Teaming
- Adversarial Prompts
Open coursework - StrategyIntermediateNew
Migrating a 40-Person SaaS Scale-Up to Cloud-Native Architecture
You are to create a detailed migration plan for TaskFlow. The plan must include breaking the monolith into at least 4 microservices, containerizing them with Docker, orchestrati…
- Microservices
- Docker
- Kubernetes
Big Data and Cloud Technologies - Browse challenges
Explore role
Product Manager
Ship product that solves real user problems. Combine user research, prototyping, and stakeholder alignment to turn ambiguous briefs into measurable wins — the role at the centre of modern software teams.
- CodeBeginnerNew
Build a Real-Time Operations Wall Display for a Logistics AI Startup
You receive a websocket feed of operational events (around 20 events per second) plus a small KPI definition list (throughput per zone, late-truck count, exception queue depth, …
- Realtime Visualization
- Glanceability
- D3
Data Visualization - AnalysisIntermediateNew
Exploration Strategies for a Recommendation Bandit
You receive 60 days of anonymized impression/click logs covering around 200 content items and user features (cohort, listening history bucket). Build a contextual-bandit simulat…
- Contextual Bandits
- Thompson Sampling
- Ucb
Reinforcement Learning - ResearchSeniorNew
Benchmark Long-Context Architectures on a Legal-Doc Retrieval Task
You receive a public legal-QA dataset (e.g., LongBench's legal split or similar) filtered to documents over 50,000 tokens. Implement or wrap 3 architectures: a sliding-window Tr…
- Long Context Architectures
- State Space Models
- Transformers
Advanced Deep Learning - CodeIntermediateNew
Build an Evaluation Harness for an Internal LLM Assistant
You will design and implement an evaluation harness in Python that runs four test suites: (1) helpfulness (LLM-as-judge with rubric), (2) factual grounding (compare cited source…
- LLM Evaluation
- LLM As Judge
- Prompt Injection Testing
Large Language Models Build a verifiable portfolio.
Submissions become evidence. Reviewers with shipping experience score against a rubric; the result becomes a credential anyone can verify.
Why Ewance
- AnalysisBeginnerNew
Build a Topic-Modeling Pipeline for Citizen Feedback
Take the 60,000 comments (anonymized). Build a BERTopic pipeline with multilingual sentence embeddings (Catalan + Spanish + occasional English). Tune number-of-topics via topic-…
- Topic Modeling
- Bertopic
- Multilingual NLP
Open coursework - CodeIntermediateNew
Train a VAE for Synthetic Tabular Data at a Healthtech Startup
You receive a synthetic-but-realistic clinical-trial table (around 50,000 patients, 35 columns, mixed continuous and categorical). Train a tabular VAE (or TVAE/CTGAN as alternat…
- Vae
- Tabular Generation
- Synthetic Data
Deep Generative Models - AnalysisBeginnerNew
Audit a Hiring-Screen Classifier for Fairness Across Cohorts
You receive the classifier as a black-box API and a synthetic-but-realistic dataset of 8,000 CVs with imputed demographic proxies (gender, age band, regional cluster) and labele…
- Fairness Evaluation
- Disparate Impact
- Audit Methodology
Trustworthy AI, Robustness, and Safety - AnalysisBeginnerNew
Diagnose Query Failures in an E-Commerce Search Box
You receive 6 months of anonymized query logs (~480 million rows): query string, language hint, results-shown count, top-3 product clicks, and add-to-cart events. Build a notebo…
- Query Log Analysis
- Clustering
- Ir Failure Analysis
Open coursework - CodeSeniorNew
Train a Manipulation Policy for Bin Picking with Imitation Learning
You receive a dataset of 500 teleop trajectories on the in-distribution part plus a held-out simulation environment with a never-seen part. Train an imitation-learning policy (D…
- Imitation Learning
- Manipulation
- Diffusion Policy
Advanced Robotics - CodeIntermediateNew
Fine-Tune a Transformer for Customer-Support Triage at an Enterprise AI Vendor
You receive 240,000 labeled support tickets across 14 queues, with English, Bahasa Indonesia, and Tagalog. Fine-tune a multilingual transformer encoder (XLM-RoBERTa-base is a st…
- Transformers
- Fine Tuning
- Multilingual NLP
Deep Learning - CodeSeniorNew
Multilingual RAG for a European Customer-Support Knowledge Base
You receive 6,000 documents in 4 languages (mix of FAQs, parts catalogs, repair procedures) plus 120 labeled queries (30 per language) with gold source documents. Build a multil…
- Multilingual RAG
- Cross Lingual Retrieval
- Multilingual Embeddings
Open coursework - AnalysisIntermediateNew
Catastrophic-Forgetting Audit on a Domain Fine-Tune
You receive the fine-tuned 7B chemistry model and its base, plus a benchmark basket (MMLU subset, GSM8K, IFEval, a small instruction-following set). Run all 4 benchmarks on both…
- Catastrophic Forgetting
- LLM Evaluation
- Fine Tuning
Fine-Tuning Large Language Models - ResearchIntermediateNew
Reproduce a Vision-Model Paper Under a Reproducibility Standard
Pick a vision-model paper from CVPR or NeurIPS 2024-2025 with publicly available code and a manageable compute footprint (single-GPU under 24 hours). Reproduce the headline metr…
- Reproducibility
- Experimental Design
- Model Evaluation
AI Measurement and Evaluation - CodeIntermediateNew
Multi-View Pose Estimation for a Sports-Analytics Startup
Use the publicly-released SoccerNet or a synthetic 4-view dataset (you can render with Unity or use a provided one). Implement a 2D pose estimator per view (HRNet or YOLOv8-pose…
- Pose Estimation
- Multi View Geometry
- 3d Reconstruction
Open coursework - StrategyIntermediateNew
Design a PETs Strategy for an EU AI Act Use Case
Map the underwriting use case to applicable PETs across the data-lifecycle stages (training, evaluation, inference, monitoring). For each, document: privacy property gained, acc…
- Pets Strategy
- Differential Privacy
- Federated Learning
Open coursework - CodeIntermediateNew
Finetune a Diffusion Model for Sustainable-Fashion Mockups
You receive 1,200 product photos with paired captions and the brand's style guide. Fine-tune a Stable-Diffusion-class base model with LoRA (Low-Rank Adaptation, a parameter-effi…
- Diffusion Models
- Lora Finetuning
- Pytorch
Open coursework
How it works
From brief to credential, in six steps.
Step 01
Browse challenges aligned to your studies.
Step 02
Accept the one that fits your goals.
Step 03
Work through it with AI Copilot guidance.
Step 04
Submit for structured evaluation.
Step 05
Earn a verified credential.
Step 06
Add it to LinkedIn with one click.
Industry teams behind a decade of practitioner briefs
Hiring from this pool?
Sponsor a challenge and meet candidates through actual work.
Industry teams can shape briefs around the skills they hire for, then evaluate students on rubric-scored deliverables — not resumes.