AI & Data
Deep Learning Challenges
Deep Learning challenges put you inside the work of building models that learn from raw data. You'll develop skills in Neural Networks and Feedforward Networks, apply Data Augmentation, and train models in PyTorch or TensorFlow alongside Reinforcement Learning fundamentals.
From there you'll handle the harder edges — Transformer architecture, Attention mechanisms, Custom architecture design, and Distributed training — working with PyTorch Lightning / Hugging Face Trainer, JAX research patterns, and Ablation study design the way research teams actually do. Each challenge you solve earns a verified credential you can share with recruiters.
- AnalysisIntermediateNew
Structured Prediction for Insurance Claim Triage
You receive 18,000 historical claims with text, attachments-count, claim amount, customer tenure, and the ground-truth final routing bucket. Train a structured classifier (e.g.,…
- Structured Prediction
- Multi Class Classification
- Model Evaluation
Advanced Machine Learning - ResearchSeniorNew
Embodied Visual Reasoning for a Warehouse Pick Assistant
Use an embodied simulator (Habitat 3.0 or Isaac Sim — pick one and justify) to render 300 cluttered-bin scenarios with a target item label. For each scenario, build two reasonin…
- Embodied Vision
- Vision Language Models
- Visual Reasoning
Visual Intelligence and Visual Reasoning - 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
Privacy-Preserving Machine Learning - ResearchSeniorNew
Open-Vocabulary Segmentation Benchmark for a Robotics R&D Lab
Use a curated 200-image household scene set (publicly-available HM3D renderings or COCO + a handful of household prompts). Benchmark 3 open-vocabulary segmentation models: SAM +…
- Open Vocabulary Segmentation
- Vision Language Models
- Benchmarking
Computer Vision Practice your coursework on real scenarios.
Every challenge is shaped from real industry context — not generic exercises. The work mirrors what your degree prepares you for.
Why Ewance
- CodeIntermediateNew
Build a GAN-Based Defect Generator for a Hardware Manufacturing Line
You receive around 60,000 good-unit images and around 380 defective-unit images across 4 defect classes. Train a class-conditional GAN (StyleGAN2-ADA or a smaller alternative fo…
- Gans
- Class Conditional Generation
- Data Augmentation
Deep Generative Models - CodeIntermediateNew
Scene-Graph Generation for Retail Shelf Audits
You receive 1,500 labeled shelf photos (anonymized product crops, bounding boxes, and ~12 relation types). Build a pipeline that, for a new shelf photo, outputs (a) detected pro…
- Scene Graph Generation
- Object Detection
- Relation Prediction
Visual Intelligence and Visual Reasoning - CodeBeginnerNew
Compare MDP Solvers for a Smart-Grid Battery Dispatch Pilot
Model home-battery dispatch as a finite MDP: state is (state-of-charge, hour-of-day, current price tier), actions are charge/hold/discharge with realistic efficiency losses, tra…
- Markov Decision Processes
- Value Iteration
- Policy Iteration
Artificial Intelligence: Principles and Techniques - CodeIntermediateNew
Generate Synthetic Tabular Data with Privacy Guarantees
Implement DP synthetic data generation: either DP-CTGAN, PATE-GAN, or a marginal-based DP method like PrivBayes / MWEM. Train on the real dataset (around 200,000 transactions, 1…
- Synthetic Data
- Differential Privacy
- Generative Models
Privacy-Preserving Machine Learning - 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.
- CodeIntermediateNew
Fine-Tune a Sequence-to-Sequence Model for Code-Doc Generation
Take a small base model (CodeT5+ or a distilled CodeLlama-Instruct). Build the dataset by mining around 8,000 high-quality function-docstring pairs from permissively-licensed Py…
- Seq2seq
- Hugging Face Transformers
- Fine Tuning
Neural Networks for NLP - 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 - 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
Computer Vision - ResearchSeniorNew
Certify Robustness for a Medical-Imaging Classifier
You receive the classifier (a PyTorch ResNet variant) and a 4,000-image labeled validation slice. Apply randomized smoothing (Cohen et al.) at sigma in {0.25, 0.5, 1.0}. Report …
- Certified Robustness
- Randomized Smoothing
- Formal Verification
Trustworthy AI, Robustness, and Safety 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
- ResearchSeniorNew
Investigate Why Our Generative Model Memorizes Training Data
Pick a small open-source diffusion model (e.g., a Stable-Diffusion-class community model trained on LAION-subset). Reproduce a published membership-inference + extraction probe …
- Generative Models
- Memorization Analysis
- Differential Privacy
Advanced Deep Learning - CodeIntermediateNew
Extractive QA on Clinical Trial Protocols
You receive 500 anonymized protocol PDFs (already OCR-ed to text) and 1,200 labeled question-answer pairs where each answer is an exact text span. Build an extractive QA system:…
- Extractive Qa
- Reading Comprehension
- Model Finetuning
Question Answering and Conversational Systems - ResearchSeniorNew
Quantify Sim-to-Real Gap for a Warehouse Manipulation Policy
You receive a trained pick-and-place policy (PyTorch), the simulation env (Isaac Lab), and access to a real-arm rig (or recorded teleop episodes if hardware is unavailable). Def…
- Sim To Real
- Manipulation
- Experimental Design
Robot Perception and Autonomy - ResearchIntermediateNew
Fine-Tune a Vision-Language Model for Image Captioning
Take BLIP-2 or LLaVA-1.6 as the base. Fine-tune (LoRA is fine) on a 4,000-image accessibility-curated dataset where each image has a useful caption written by a low-vision-exper…
- Vision Language Models
- Fine Tuning
- Pytorch Or Tensorflow
Multimodal Machine Learning - ResearchSeniorNew
Trajectory Prediction Model for Urban Robotaxis
Use the Argoverse 2 motion-forecasting dataset (open access). Train an LSTM baseline + a transformer challenger (e.g., a small Wayformer or HiVT). Evaluate on minADE/minFDE (min…
- Trajectory Prediction
- Transformer Models
- Evaluation
AI for Autonomous Vehicles - ResearchIntermediateNew
Multi-Task Learning for a Healthtech Triage Model
You receive 40,000 anonymized de-identified intake-form records with two labels: urgency tier (4 classes) and routed sub-specialty (12 classes). Train (1) two independent classi…
- Multi Task Learning
- Transfer Learning
- Hugging Face Transformers
Meta-Learning, Transfer Learning, and Multi-Task Learning - CodeSeniorNew
Video Action Recognition for a Retail Loss-Prevention Startup
Use a public action-recognition dataset (UCF101 + a small curated retail-action subset; the latter is provided synthetic or you can label 50 short clips). Fine-tune a small back…
- Video Understanding
- Action Recognition
- Transfer Learning
Computer Vision - CodeIntermediateNew
Build a Speaker-Diarization Pipeline for a Legal-Tech Startup
You receive 20 hours of de-identified hearing audio with ground-truth speaker labels (4 speaker classes per hearing). Build a speaker-diarization pipeline (pyannote-audio or sim…
- Speaker Diarization
- Speech Recognition
- Pyannote
Speech Recognition and Spoken Language Processing - CodeIntermediateNew
Build a Sequence Model for Sign-Language Word Recognition
You receive about 12,000 short (1-3s) webcam clips covering a 50-word vocabulary, with body+hand pose features pre-extracted (e.g., MediaPipe Holistic landmarks per frame). Buil…
- Sequence Models
- Hugging Face Transformers
- Pose Estimation
Machine Perception - CodeIntermediateNew
Train a Multimodal Classifier for Medical Triage
Pick a fusion architecture (early fusion via cross-attention, late fusion via score combination, or a unified multimodal encoder like FLAVA/CoCa). Train on the 14,000 pairs with…
- Multimodal Fusion
- Cross Attention
- Pytorch Or Tensorflow
Multimodal Machine Learning - CodeIntermediateNew
Quantize a CNN for Battery-Powered Wildlife Cameras at a Climate Nonprofit
You receive an FP32 CNN (MobileNetV2 fine-tuned to 22 species, around 13 MB) and a hold-out test set of 4,000 images. Quantize to int8 (post-training quantization first, then qu…
- Quantization
- Qat
- Edge Deployment
Deep Learning - CodeBeginnerNew
End-to-End Lane Following on a Donkeycar Platform
Use the public Donkeycar Tub dataset (or collect about 30 minutes of driving on the simulator). Train a CNN-policy baseline (the Donkeycar default architecture is fine) that pre…
- End To End Learning
- Imitation Learning
- Pytorch Or Tensorflow
AI for Autonomous Vehicles
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.