Machine Learning Engineer (Irrigation Auto-Scheduling)

Machine Learning Engineer (Decision Systems & Automation)
Remote role with partial US business-hours overlap required.

?? Your application email subject line MUST read:
“I actually read the automation instructions”

Who We Are

We are Buzz Insider, partnering with leading B2B companies to build intelligent, reliable systems that reduce manual work and improve data integrity. We value people who ship thoughtfully, explain clearly, and elevate the bar.

Core Responsibilities (General)

Problem framing & metric design: Turn business goals into clear modeling objectives, success metrics (MAE/precision/latency), and guardrails.

Data ingestion & normalization: Build lightweight pipelines from APIs/files/DBs, validate schemas, handle missing/outliers, and persist tidy tables.

Feature engineering & baselines: Create transparent rule-based baselines and analytical heuristics to establish a reliable starting point.

Modeling & residual learning: Train small, practical tabular models (e.g., ridge/GBDT) to correct baseline bias using recent outcomes; track versions and reproducibility.

Decision logic & policy layer: Add thresholding, hysteresis, fallbacks, and overrides that convert predictions into auditable actions.

Anomaly/change-point detection: Detect unexpected behavior in time-series/logs (CUSUM/change-point tests) and produce actionable i ---------- /alerts.

Constraint-aware scheduling/allocation: Build simple packers/optimizers that place tasks within allowed windows while respecting capacity/SLA limits.

Observability & evaluation: Emit decision logs (inputs ? features ? outputs ? reasons) and provide evaluation scripts & lightweight dashboards.

Integrations & automation: Expose/consume REST/webhooks; trigger actions in external systems in dev/sandbox environments.

Quality & ops hygiene: Add unit/integration tests, structured logging, env-based configs, reproducible CLIs, and basic CI.

Documentation & runbooks: Keep concise docs: data contracts, assumptions, decision policies, troubleshooting, and handoff notes.

Privacy & security: Manage credentials safely, redact sensitive data in logs, and apply least-privilege access.

Skills

Core technical

Python (pandas, numpy, scikit-learn); bonus: LightGBM/XGBoost, statsmodels, pydantic/typing

SQL/Postgres (joins, window functions, partitions), basic performance tuning

APIs & services: FastAPI/Flask, REST/webhooks, auth (API keys/OAuth), pagination/rate limits

Time-series & tabular modeling: baselines, residual learners, forecasting, anomaly/change-point detection

Scheduling/optimization: heuristics/packers; bonus: OR-Tools/CP-SAT

Testing & quality: pytest, deterministic seeds, “golden” datasets; (nice-to-have: Great Expectations)

MLOps hygiene: reproducible CLIs/scripts, experiment tracking (MLflow), model/version artifacts

DevOps: Docker, Git/GitHub, CI (GitHub Actions), logging/metrics (structured logs, Prometheus/Grafana)

Cloud & security basics: env configs, secrets management, least-privilege/IAM, PII redaction

Professional

Clear written docs (schemas, assumptions, runbooks), PR discipline, async collaboration

Ability to translate business goals ? measurable metrics/guardrails and explain model behavior in plain language

Capabilities (What you’ll own)

Data contracts & pipelines: Define minimal schemas; build API/file loaders with validation and backfills; document lineage & freshness.

Decision engines: Ship transparent baselines; upgrade with lightweight ML to reduce bias/variance; maintain explainability.

Detection & reliability: Implement anomaly/change-point alerts; create i ---------- records and triage rules; track data health & drift.

Constraint-aware execution: Provide packers/schedulers that respect time windows, capacity, and concurrency; include dry-run/explain/override modes.

Observability & evaluation: Maintain evaluation scripts/dashboards (accuracy/MAE, coverage, policy compliance, latency).

Integrations & delivery: Expose inference endpoints; consume external services; package with Docker; wire basic CI; deliver “operate this” runbooks.

Governance & safety: Version data/code/models; keep rollback plans; handle secrets correctly; log responsibly; enforce access scopes.

Candidate Profile

You are methodical, detail-oriented, and process-driven. You think in systems and take pride in shipping small, correct components that add up to reliable automation.

Traits we value

Logical and structured thinker

Strong problem solver

Self-driven and resourceful

Attentive to detail

Motivated by efficiency and scale

What energizes you

Turning messy signals into clear decisions (minutes, schedules, shutoffs, alerts)

Shipping baselines fast, then improving with data

Explaining model decisions in plain language

To Apply

Send your application to ---------- emails sent to this address will be considered.

Include:

Resume (PDF link, format: Lastname.Firstname.ML.pdf)

Portfolio – 2–3 examples of ML/modeling projects (notebooks, repos, or screenshots)

Animal Question: “If you were an animal, what would you be and why?”

Experience: “How many years have you built data/ML systems?”

A brief description of the most complex data/ML system you’ve built and how you overcame challenges

Your available hours in US Eastern Time

?? Format your submission professionally — attention to detail will be evaluated. Most applicants will miss these instructions and be disqualified. Those who follow them will be shortlisted.