# Would someone be kind enough to review an MLOps platform portfolio project? — retained record

> Retained reddit record from reddit:MLOps.

Canonical: https://highsignal.app/data/records/aaeeebeb37c2681c

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:MLOps)  published 2026-09-25 08:12 ingested 2026-09-26 02:36

[original source ↗](https://www.reddit.com/r/mlops/comments/1wpqmpp/would_someone_be_kind_enough_to_review_an_mlops/)

## retained text



```
I recently concluded an MLOps portfolio project that I want to use to find a job. It is fully documented with writeups and diagrams (that are well-written by myself) to explain the entire platform and covers everything from architectural decisions to the data science problem to the model (fine-tuned Hermes 4 -14b) to all the workflows. Please reach out to me privately so I can send you the github link, or let me know what you think about the following extract from my resume.

**MLOps platform with CI/CD**  
• Built MLOps platform using Terraform, Kubernetes, MLRun (Python) with pipelines for training (QLoRA fine-tuning, PyTorch), deploying, and monitoring a 14B LLM.

• Serves model that extracts legal risks/restrictions/obligations from multi-page contracts with source attribution into JSON data (for human verification) with Sagemaker and vLLM on AWS.

• Utilises data/model (and prompt) registries, experiment tracking, drift detection rollback, canary deployment, and continuous training. Maintained service levels with multi-GPU training/serving, autoscaling, quantisation, continuous batching, and performance benchmarks.
```



document fetched 2026-09-26 00:28 · canonical https://reddit.com/r/mlops/comments/1wpqmpp/would_someone_be_kind_enough_to_review_an_mlops
