Use Cases
See how NIKI helps developers in different domains
🤖 Machine Learning
ML Training Jobs
Training deep learning models takes hours or days. NIKI understands ML terminology and extracts the metrics you care about.
- ✓ Loss and accuracy tracking
- ✓ Best checkpoint identification
- ✓ Early stopping detection
- ✓ GPU memory monitoring
$
niki watch
-- python train.py --epochs 1000
# Notification:
✓ Training Complete (2h 34m)
Loss: 0.0023 (↓ from 0.847)
Accuracy: 98.7%
Best checkpoint: epoch 847
Accuracy: 98.7%
Best checkpoint: epoch 847
$
niki watch
-- cellranger count \
--id=sample_01 \
--transcriptome=refdata
# Notification:
✓ CellRanger Complete (4h 12m)
Cells detected: 12,847
Median genes/cell: 2,341
Mapping rate: 94.2%
Median genes/cell: 2,341
Mapping rate: 94.2%
🧬 Bioinformatics
Bioinformatics Pipelines
scRNA-seq, alignments, variant calling - these take hours. NIKI understands bioinformatics tools and QC metrics.
- ✓ Cell Ranger / STAR / BWA support
- ✓ QC metric extraction
- ✓ Scanpy / Seurat integration
- ✓ Snakemake / Nextflow pipelines
📊 Data Engineering
Data Processing & ETL
Large-scale data transformations, batch processing, and ETL jobs. Know when your data is ready.
- ✓ Row count tracking
- ✓ Error rate monitoring
- ✓ Cloud storage integration
- ✓ Airflow / dbt awareness
$
niki watch
-- python etl.py \
--input s3://bucket/raw/ \
--output s3://bucket/processed/
# Notification:
✓ ETL Complete (1h 45m)
Rows processed: 47,382,193
Invalid rows: 1,247 (0.003%)
Output: 12.4 GB
Invalid rows: 1,247 (0.003%)
Output: 12.4 GB
$
niki attach
12345 --name "GPU job"
# Attach to Slurm job
# Notification:
✓ Slurm Job Complete (8h 23m)
Job ID: 12345
Node: gpu-0042
GPU peak: 31.2 GB / 40 GB
Node: gpu-0042
GPU peak: 31.2 GB / 40 GB
🖥️ High Performance Computing
HPC & Slurm Jobs
Running jobs on clusters? Attach to existing Slurm jobs and get notified when they finish.
- ✓ Attach to running jobs
- ✓ GPU memory tracking
- ✓ Node information
- ✓ Works through SSH