Variant 01 · Most popular
Data Scientist (Mid-Level)
The default template for data scientists with 3-5 years of experience. Balances modeling depth, experimentation skill, and demonstrated business impact across 2-3 employers.
3-5 years expIC2 / IC3$130K-$175K range
Priya Kapoor
Data Scientist
Data Scientist with 4 years building production ML and experimentation programs. Shipped models touching 8M weekly users. Recovered $4.2M annual revenue via earlier churn detection.
Data Scientist · Lyft · 2021 — Present
- Built driver-churn prediction model (XGBoost, Snowflake, dbt) serving 8M weekly active riders; identified at-risk users 14 days earlier than legacy heuristic, recovering $4.2M annually.
- Designed pricing-test platform (Python, Airflow) used by 40+ analysts; cut experiment design-to-launch time from 11 days to 3.
- Partnered with product and ops on driver-incentive redesign, growing weekly active drivers 9% in two pilot regions.
How it differs: Focus on ownership of models end-to-end, not just contributing to one. Include 1-2 cross-functional collaborations with product or marketing. Skills section should be deep in Python + SQL + 1 ML framework, not broad across 5.
Variant 02 · For new grads
Entry-Level / Junior Data Scientist
For new grads, bootcamp graduates, and DS with under 2 years of experience. Leads with projects and education when work history is thin, and treats internships as full jobs.
0-2 years expNew grad$95K-$135K range
Daniel Park
Data Science Intern → Full-Time
Statistics MS (3.9 GPA, Carnegie Mellon) with summer internship at Capital One Risk. Kaggle Competitions Expert (top 2% in 2 NLP comps). Strong in Python, SQL, scikit-learn, PyTorch.
Capstone: Credit-Risk Default Modeling (Project) · 2024
- Built default-prediction model on Lending Club dataset (XGBoost, Python, 2.2M rows); achieved 0.91 AUC on holdout, presented to industry panel of 3 reviewers from JPMorgan and Goldman Sachs.
- Published 'streamfeature' on GitHub (1,800+ stars), a Python library for streaming feature aggregation built on Apache Beam, adopted in production by 2 mid-stage startups.
How it differs: Project section comes before experience. Education listed in full (GPA if 3.5+, relevant coursework, capstone, thesis). Include 2-3 personal/open-source/Kaggle projects with metrics (leaderboard rank, stars, downloads). Internships count as full jobs.
Variant 03 · For IC4-IC5
Senior Data Scientist
For data scientists with 6+ years targeting senior roles. Emphasizes scope of business impact, technical leadership without a manager title, and influence across product teams.
6-10 years expIC4 / IC5$180K-$280K range
Wei Zhang
Senior Data Scientist
Senior Data Scientist with 8 years owning marketplace ML at DoorDash. Led 5-DS team rebuilding matching model. Set experimentation standards across 3 product squads. Mentored 9 data scientists, 3 promotions.
Senior Data Scientist · DoorDash · 2021 — Present
- Led 5-DS team rebuilding marketplace matching model (PyTorch, real-time inference on AWS); shipped to all 9M monthly customers, +6.2% completion rate worth $90M annualized.
- Set team's experimentation standards across 3 product squads; introduced sequential testing and CUPED variance reduction, cutting required sample sizes 40% and accelerating shipping cadence.
- Authored DS hiring rubric; interviewed 60+ candidates, hired 4 (3 still on team 2+ years later).
How it differs: Bullets should describe systems and programs you owned, engineers you mentored, and standards you set — not just models you shipped. Include 1-2 'led X DS' or 'set experimentation direction for Y org' statements. Track record of leading something is what separates Mid from Senior.
Variant 04 · Production-ML focus
Machine Learning Engineer / Production ML
For data scientists who own ML systems in production — model serving, monitoring, retraining pipelines. Skills lean toward engineering: Python, distributed systems, Kubernetes, MLOps.
3-8 years expProduction ML$150K-$290K range
Marcus Lee
Machine Learning Engineer
ML Engineer with 6 years owning production ML systems at Pinterest. Built recommendation infrastructure serving 7.8B daily predictions at <50ms p99. Stack: PyTorch, Triton, Kubernetes, MLflow.
Machine Learning Engineer · Pinterest · 2021 — Present
- Owned end-to-end recommendation infrastructure (PyTorch, Triton, Kubernetes) serving 7.8B daily predictions at <50ms p99 latency across 480M monthly users.
- Built model-monitoring framework (drift detection + automatic retraining triggers) adopted across 14 production models; prevented 3 silent regressions in year 1.
- Cut model deployment cycle from 2 weeks to 11 hours by rebuilding CI/CD on Kubeflow Pipelines.
How it differs: Drop the analyst framing entirely. Lead with infrastructure: latency, throughput, uptime, retrain cadence. Include MLOps stack (Kubeflow, MLflow, Triton, Vertex AI). Hiring managers screen for 'can run a model in prod for 12 months without it silently breaking' — show that.
Variant 05 · Experimentation focus
Product / Experimentation Data Scientist
For DS focused on product analytics, causal inference, and A/B testing. Closer to PMs and product orgs than to ML platform teams; less modeling code, more experiment design and decision support.
3-7 years expCausal + experimentation$140K-$220K range
Sofia Reyes
Product Data Scientist
Product Data Scientist with 5 years at Notion and Stripe. Designed and analyzed 60+ A/B tests on onboarding funnel. Built causal-inference toolkit replacing $260K/year in external consulting.
Product Data Scientist · Notion · 2022 — Present
- Designed and analyzed 60+ A/B tests on onboarding funnel; recommendations shipped to 8M users contributing +4.1% W4 retention worth $11M ARR.
- Built causal-inference toolkit (synthetic control + diff-in-diff) for go-to-market evaluation; replaced consultant-built quasi-experiments saving $260K/year in external fees.
How it differs: Lead with experiments shipped, not models trained. Bullets should describe decisions you influenced, not architectures you wrote. Include causal-inference methods (synthetic control, diff-in-diff, instrumental variables) and stakeholder names you partnered with (PMs, designers, GMs).
