Python 2 - usable

Python 3 - preferred

Natural Language Processing

Machine Translation

Regression, classification, anomaly detection, recommendation, ...

Model evaluation and interpretation

Notebooks (e.g. Jupyter, IPython)

Data acquisition (e.g. SQL, web scraping, OCR, writing custom parsers for weird internal binary formats)

EDA (e.g. Pandas, DuckDB, Spark, or even using built-in csv)

Visualizations (e.g. Seaborn, Matplotlib, Plotly)

Baseline models (e.g. anything in scikit-learn, xgboost/etc, few-shot prompting for LLMs)

Statistical Analysis (e.g. scipy, numpy, scikit-learn)

Regular expressions, including the obscure bits

Microsoft Powerpoint, Word, Excel, and Access when necessary

Self-built configuration management tool

Dockerfiles, docker-compose, K8s and kubectl, Rancher, Helm

Terraform (mostly for Keycloak)

OAuth2, OIDC, JWTs, HTTP-only cookies, 2FA, OPA, Envoy, Rego

Keycloak, Active Directory

Kong, DecK, Konga, that one open-sourced OIDC plugin

Linux terminal, but mostly Bash

Hazelcast, Memcached, Postgres, Oracle database, OpenSearch (ELK), Directus

OpenTelemetry instrumentation

Using Jira, Confluence, Mattermost

Setting up monitoring, health checks, and alerting

SAST, type checking for python

Flask, FastAPI, SQLAlchemy, SQLModel, Streamlit

Git, Gitlab CI

Networking (e.g. 40G QSFP, 10GBASE-T, VLANs, some knowledge of edge switches and core routers)

Power (e.g. PDUs, AC phases, earthing, power loading, CEE and IEC plugs)

Cooling (e.g. ASHRAE thermal guidelines, hot/cold zones, and airflow

Racking (e.g. standard rack sizes, capacities, lock types)

Installing the physical servers, wiring the servers, RMA-ing some servers 🙄

Setting up the servers (e.g. RAID on HBA controllers, networking, IPMI, VMWare)

HTML, CSS, Javascript, Bootstrap, and jQuery

Streamlit, FastAPI, Flask