Reproducible models with visible operational health
MLOps Engineer
My MLOps direction connects model experiments to the conditions that make them dependable after training. The in-progress anomaly pipeline spans versioned data, tracked experiments, drift signals, container delivery, autoscaling, rollback planning, and service telemetry rather than ending at a notebook metric.
Why this direction fits
I am building the habits that make model work repeatable and operable: version inputs, register experiments, package inference, watch drift and latency, and define recovery before deployment.
Current evidence
BCA in Information Technology, Oracle Data Science Professional, IIT Dharwad-aligned deployment training, and two MLOps-relevant in-progress systems.
01 / CAPABILITY MAP
How I would contribute
Reproducibility
DVC datasets, MLflow experiments, model registration, and comparable training runs.
Deployment
Docker packaging, Kubernetes rolling updates, HPA, and planned rollback triggers.
Observability
Prometheus metrics, Grafana views, health signals, latency, volume, and drift visibility.
Lifecycle
Automated retraining direction, evaluation gates, model versioning, and safe release practices.
02 / VERIFIED SKILLS
Role-aligned toolkit
- Python 3.13
- XGBoost
- PyTorch
- MLflow
- DVC
- Docker
- Kubernetes
- Prometheus
- Grafana
- GitHub Actions
03 / SELECTED WORK
Evidence for this role
Production MLOps Anomaly Detection & Deployment System
A versioned anomaly pipeline with experiment tracking, drift monitoring, autoscaling, rollback triggers, and observable serving.
- Target: automated drift visibility
- Target: safe rolling updates with rollback
Multimodal Defect & Threat Inspection System
A multimodal inspection pipeline combining lightweight vision models, signal metadata, optimized inference, and anomaly records.
- Target: 42 FPS edge inference
- Target: 94.2% mAP across five categories
High-Throughput Asynchronous AI Gateway & Task Engine
An asynchronous gateway with distributed tasks, provider fallback, token-bucket limits, deduplication, and auditable usage.
- Target: 250+ concurrent requests
- Target: sub-50ms queue latency · >80% coverage
04 / TAILORED RESUME
Resume for this role
MLOps Engineer
This version will foreground the experience and project evidence most relevant to this hiring path.
- —Experiment and dataset reproducibility
- —Containerized inference and deployment safety
- —Drift, service health, and model observability