industry-transition

name	industry-transition
description	Translate a PhD into an industry resume, navigate technical screens and system design for ML roles, and compare biotech, big-pharma, ML-for-bio, science writing, and consulting paths. Includes startup vs established company trade-offs. Use when a PhD student or postdoc is preparing for a job search outside academia.
license	MIT

PhD to Industry Career Transition

Hard rules

1. No fabricated citations. Every cited work must resolve to a verifiable
2. No claim without provenance. Every quantitative or factual claim
3. No silent failure. Every script invocation, API call, or tool use must declare its exit status and what to do on non-zero. A skill that silently swallows errors is a violation.

When to use

• PhD student or postdoc considering a non-academic job search.
• A trainee whose timeline has shifted and academic positions are not the primary target."
• Tailoring a resume for a specific industry role (e.g., "scientist II" at a biotech vs "research engineer" at an ML-for-bio startup).
• Preparing for technical screens, take-home assignments, and on-sites in industry interviews.
• Building a target list of companies by sector and stage (early-stage startup vs late-stage public company).

When NOT to use

• Academic job market — see ors-career-navigation-faculty-interview and ors-career-navigation-academic-cv.
• Salary negotiation in detail (covered briefly here, in depth in ors-career-navigation-negotiation).
• Career switch for an industry veteran; the framing here is for an academic transitioning to industry.
• Government / national-lab transition (overlaps with industry, but federal hiring has its own process — USAJobs, KSAOs, GS scale).

Prerequisites

• A current CV (see ors-career-navigation-academic-cv).
• A 1-2 page summary of research accomplishments with quantified impact.
• A list of 3-5 projects you can talk about in depth (the "story bank").
• Public information on target companies (careers pages, Glassdoor, LinkedIn, Levels.fyi for compensation).
• A network contact at one or more target companies (alumni, conference contacts, Twitter/X, the company's "people" page).

Core workflow

1. Decide which sector to target

The sector decision shapes the resume, the network, and the interview prep. Five common sectors for biomedical PhDs:

Biotech (early-stage, often Series A-C). Research scientist at a venture-backed startup. Pros: high ownership, broad scope, equity, mission-driven; cons: lower base salary, less stability, sometimes long hours. Roles: research scientist, bioinformatics scientist, ML scientist, platform scientist.

Big pharma (Genentech, Regeneron, Amgen, Vertex, and similar large-cap biotechs). Scientist I/II in a therapeutic area or a platform group. Pros: structured ladder, mentorship, benefits, often hybrid-able; cons: more siloed, slower to publish, can be more hierarchical. Roles: scientist, senior scientist, principal scientist (track); computational biology, biostatistics, bioinformatics, genomics, clinical genomics.

ML-for-bio / computational biology in tech-adjacent companies. Research scientist or applied scientist at a company that builds ML tools for biology (Alphabet's biology efforts, large tech companies with bio/health divisions, and a growing set of startups). Pros: state-of-the-art ML, high compensation, strong engineering culture; cons: role can be detached from wet-lab validation, more competitive hiring. Roles: research scientist, applied scientist, ML engineer.

Science writing and communication. Staff writer at a journal (Nature, Science, Cell Press trade press), a popular outlet (Quanta, STAT, The Atlantic science), or an in-house writer at a company. Pros: leverages PhD in a different way, often remote-friendly; cons: typically lower compensation, fewer roles, requires a portfolio. Roles: science writer, science editor, communications lead.

Consulting (life-sciences strategy, biotech consulting, general management consulting). Consultant at a firm (McKinsey, BCG, Bain in general management; ZS Associates, Charles River Associates, or boutique life-sciences strategy firms for life-sciences focus). Pros: high compensation, broad exposure, fast network growth; cons: travel, client-facing hours, less hands-on science. Roles: associate consultant, consultant, manager.

Other sectors that overlap with these and are worth considering: medical affairs, clinical development, regulatory affairs, venture capital (research associate roles), patent law (with a JD), science policy, government science (NIH, FDA, CDC, NSF).

2. Translate the CV to a one or two-page industry resume

Industry resumes are short, dense, and accomplishment-oriented. The academic CV is long, comprehensive, and publication-oriented. Translation rules:

• Length: 1 page for candidates with less than 10 years of post-PhD experience; up to 2 pages for senior postdocs or staff scientists.
• Top of resume: name, email, phone, LinkedIn, GitHub, ORCID (optional). No photograph in US applications.
• Summary (optional, 2-3 lines): who you are, what you do, what you want.
• Education: degree, field, institution, year. Dissertation title is optional; if used, frame as a project title.
• Experience: company or lab, role, dates, location. Bullet points: lead with verbs, lead with impact, lead with metrics.
• Skills: technical skills (programming languages, frameworks, instruments, techniques) listed clearly.
• Publications: short list, "Selected publications: ..." with the most relevant 3-5. Full list goes to a separate "Publications" appendix if requested.
• Awards: short list; remove internal department awards.
• Conferences and talks: optional; industry resumes are quiet about talks.

For each bullet, the formula is: [Action verb] [what you did] [how you did it] [quantified result]. Examples:

• Academic: "Studied the role of kinase X in T cell signaling"
• Industry: "Designed and executed 12 in vivo experiments across 3 mouse models, identifying kinase X as a regulator of T cell exhaustion; results published in Cell and cited in 2 patent applications"

For ML-focused roles, swap "experiment" for "model", "dataset", "system". For science writing, lead with the byline count and outlet breadth.

3. Build a target list of companies

Build a 3-tier list per sector:

Tier	Count	Description
A — Dream	5-10	Mission, role, location, and team are all aligned
B — Strong	15-25	Most criteria met; some flexibility on location or role
C — Practice	30+	Practice interviews, build fluency, generate offers for leverage

For each company, capture: company name, sector, funding stage / size, open role, hiring manager (if findable), a contact who works there, the application deadline (or rolling), and the source of the role (LinkedIn, company site, referral).

4. Source the application

Three channels, in rough order of effectiveness for senior roles:

1. Referral — an internal contact who submits your resume into the ATS, attaches a recommendation, and signals to the hiring manager. The most reliable channel.
2. Direct application — through the company careers page. Use the company's "Open Roles" search; submit the resume; follow up on LinkedIn with the hiring manager.
3. Recruiter outbound — recruiters from search firms, in-house talent teams, or platform recruiters. The least reliable for getting seen; can still work for specific roles.

A fourth channel: published-paper citations. If your work is cited by a company, the science team may already know your name. Connect on LinkedIn with a personalized note referencing the paper.

5. Prepare for technical interviews

Technical interviews in industry vary by sector and role. Common formats:

Coding screen (1-2 hours). HackerRank / CodeSignal style: data structures, algorithms, language fluency. Prepare with LeetCode (medium difficulty is the floor for most ML or research-scientist roles). Practice in the same language you will use at work.

Take-home assignment (3-8 hours). A small project: build a model on a dataset, write a brief report, present the approach. Time-box carefully; do not over-engineer. The assignment is a test of judgment, not just technical skill.

On-site / virtual on-site (4-6 hours total). A series of 4-6 interviews:

• Technical deep-dive on your past research / projects
• Coding (similar to the screen)
• System design (for ML / engineering roles; see step 6)
• Stats / experiment design (for product or research roles)
• Behavioral (team fit, communication, conflict)
• Hiring manager / cross-functional

Prepare a "story bank" of 5-7 projects. Each story follows the STAR format (Situation, Task, Action, Result). Be ready to go deep: 5 minutes for a quick summary, 20-30 minutes for a deep-dive on a project. The interviewer will ask follow-up questions to test whether you actually did the work.

6. System design for ML roles

The ML system design interview (1 hour) tests your ability to design an end-to-end ML system, from problem framing to deployment and monitoring. A typical structure:

1. Clarify the problem (5 min): what is being predicted, what is the success metric, what are the constraints (latency, scale, cost, fairness)?
2. Data (10 min): what data is available, what are the labels, what are the leakage and bias risks?
3. Modeling (10 min): candidate model families, baselines, evaluation strategy, train/val/test split, handling class imbalance, time-based splits for time-series.
4. Features (5-10 min): feature engineering, embeddings, handling missing data, handling high-cardinality categorical features.
5. Deployment (5-10 min): online vs batch, latency, throughput, model serving, A/B test design, rollback, monitoring (data drift, label drift, performance degradation).
6. Iteration and ethics (5 min): how to improve over time, what fairness considerations apply, what regulatory requirements apply (HIPAA, GDPR).

The interviewer is testing your judgment: when to use a simple model, when a more complex one is justified, what the failure modes are. The "right" answer is usually a simple model on good data, deployed in a way that you can measure.

7. Salary negotiation and offer evaluation

See ors-career-navigation-negotiation for the full framework. For industry transitions, three points to remember:

• Cash + equity + benefits + signing bonus is the total package. Compare on total comp, not base salary alone.
• Levels.fyi, Glassdoor, and the company's SEC filings (for public companies) are public sources of compensation data. Use them to anchor a counter.
• Multiple offers create leverage. Run the search long enough to have 2-3 final-stage conversations, even if you have a strong first offer.

Do not cite specific salary numbers in your resume or interview. Reference ranges appropriate for the role and seniority, but anchor your ask in the data, not a target number.

8. Startup vs established-company trade-offs

A short, structured comparison:

Dimension	Early-stage startup	Late-stage / public company
Base salary	Lower (often below market)	Market or above
Equity	Higher (more illiquid)	Lower (often liquid)
Scope	Broad; small teams, many hats	Narrower; functional ladder
Stability	Lower (runway risk)	Higher
Mission alignment	Often very high	Variable
Mentorship	Limited; senior hires bring it	Structured; HR-driven
Work hours	Can be intense; tied to runway	More predictable
Career path	Fast, but role titles can lag	Clear ladder with promotion cycles
Publishing / open science	Variable; depends on company	Generally restricted

The "right" choice depends on your risk tolerance, your family situation, your financial cushion, and the specific role. Talk to current and former employees before deciding.

9. Common failure modes in the academic-to-industry transition

• Treating the resume as a CV: recruiters will not read a 12-page CV. Get the resume to 1-2 pages.
• Over-explaining research: the interviewer's question is "what did you do and what was the result", not "what is the deep history of your field".
• Not preparing stories: vague answers to "tell me about a project" signal a lack of self-awareness.
• Under-asking questions: candidates who do not ask about team, role, success metrics, and growth are flagged.
• Underselling technical skills: a PhD who built pipelines, models, and datasets is a stronger candidate than the resume implies.
• Not following up: a thank-you email within 24 hours is standard; a thoughtful follow-up 1-2 weeks later is rarer and more effective.

Code patterns

Resume bullet transformation (academic → industry)

# Academic bullet
"Performed ChIP-seq experiments to characterize the role of H3K27ac
in enhancer regulation during neural differentiation."

# Industry bullet (research scientist)
"Designed and executed ChIP-seq protocols (n=24 libraries) across
4 differentiation time points, identifying 1,200 dynamic enhancers
and 18 candidate regulators of neural fate; results informed 2
internal target-selection efforts and 1 external collaboration."

# Industry bullet (ML-for-bio)
"Built a deep-learning model on ChIP-seq data (PyTorch, 1.2M
training examples) to predict enhancer activity from sequence,
achieving AUROC 0.84 on a held-out chromosome split; deployed as
an internal API used by 3 teams."

Target company table

| Company | Sector | Stage | Open role | Source | Contact | Status |
|---------|--------|-------|-----------|--------|---------|--------|
| [Name]  | Biotech | Series B | Scientist II | Careers | [Name] | Applied |
| [Name]  | Big pharma | Public | Sr. Scientist | LinkedIn | [Name] | In screen |
| [Name]  | ML-for-bio | Series A | Research Engineer | Referral | [Name] | Scheduled |

System design: 1-page answer outline

# Problem
Predict [outcome] from [input], with latency [X] and accuracy [Y].

# Data
- Sources: [list]
- Labeling: [strategy, quality]
- Splits: time-based train/val/test
- Leakage: [list]

# Model
- Baseline: logistic regression on simple features
- Candidate: gradient-boosted trees or neural net
- Loss: [binary cross-entropy, etc.]
- Eval: AUROC, calibration, subgroup performance

# Features
- [list feature groups]
- Embeddings: [how generated, how updated]
- Missing data: [strategy]
- High-cardinality: [strategy]

# Deployment
- Batch or online
- Latency budget: [X ms]
- Serving: [framework]
- Monitoring: data drift, label drift, performance

# Iteration
- A/B test plan
- Rollback criteria
- Improvement roadmap

# Ethics / fairness
- [Subgroups to consider]
- [Regulatory constraints]

Common pitfalls

Pitfall	Why it fails	Fix
Submitting an academic CV	Recruiters filter by length and section headers; CV is rejected	Translate to a 1-2 page industry resume
No quantified bullets	Recruiter cannot assess impact	Use metrics: counts, effect sizes, dollar amounts, time savings
Generic objective statement	Reads as filler	Drop it; let the summary and bullets do the work
Listing every skill	Buries the relevant ones	Group skills by area; lead with the role's core stack
Not preparing for behavioral interviews	"Tell me about yourself" becomes a 10-minute ramble	Draft 2-3 versions of the opening story (60 sec, 2 min, 5 min)
Treating the take-home as a research paper	Over-invests time; signals poor prioritization	Time-box; aim for a clear, simple, well-tested solution
Misjudging the level of the role	PhD applicants often under- or over-shoot	Read the job description carefully; talk to the recruiter about level
Not researching the company	Generic questions; signals lack of interest	Read the company website, recent papers, recent product launches, leadership
Underselling PhD as "just school"	PhDs are senior candidates with deep expertise	Reframe: lead with the projects, the impact, the leadership
Quitting the academic search without a plan	6 months of anxiety, no offers	Run the academic and industry searches in parallel if the timing permits
Ignoring visa / sponsorship realities	International PhDs need to know early whether the company sponsors	Ask the recruiter in the first screen; do not assume
Negotiating on base alone	Total comp matters more	Get the full picture: base, equity, bonus, signing, relocation, benefits
Bad-mouthing academia in the interview	Signals you are running away from something	Reframe positively: what you want, not what you don't
Not following up after the on-site	A "thank you" email is standard; silence is a missed signal	Send a 1-2 paragraph thank-you within 24 hours; mention a specific conversation

Validation

A complete industry-transition plan satisfies:

• Sector decision documented; rationale written
• Resume translated to 1-2 pages; quantified bullets; tech-skills section
• Target company list with 50+ entries across tiers
• At least 2 in-flight applications / conversations in the target sector
• Story bank: 5-7 projects written in STAR format
• Coding practice: 30-50 LeetCode problems in the relevant language
• System design prep: 5-10 design problems practiced with the 1-page outline
• Salary research: levels.fyi, Glassdoor, and any company-specific data points collected
• Mock interviews completed (technical + behavioral)
• Thank-you email templates drafted
• Application tracking: simple spreadsheet or tool (Airtable, Notion) maintained
• Plan B: what if no offer by [date]?

Open alternatives

Commercial / proprietary	Open equivalent	Trade-offs
LinkedIn Recruiter	OpenResume, JSON Resume, ORCID	LinkedIn is the industry-standard network; open resume schemas help with version control
Levels.fyi (data is publicly submitted)	BLS OES data, Glassdoor community data, company SEC filings	Public data has lag; Levels.fyi is more current and granular
HackerRank / CodeSignal	LeetCode (freemium), Exercism (open), Rosalind (bioinformatics)	Exercism and Rosalind are open and bio-aware; LeetCode is the de facto standard for screens
Codility / HackerEarth	Custom assignment	Both open and proprietary have bias and fairness issues in algorithmic screening
Paid resume services	A senior mentor in your network	A trusted reviewer who knows your field is more valuable than a generic service
Paid mock interview services	Friends, mentors, recorded self-review	Both work; a structured mock with a rubric beats a generic one

References

• AAAS Science Careers: https://www.science.org/careers
• iBiology Career Skills: https://www.ibiology.org/professional-development/
• Nature Careers Turning Point column archive
• Bureau of Labor Statistics Occupational Outlook Handbook (life and physical scientists): https://www.bls.gov/ooh/
• Levels.fyi methodology: https://www.levels.fyi/comp-methodology
• Glassdoor company reviews and salary data (community-sourced)
• Bureau of Labor Statistics OES wage data (publicly available by metro area and occupation)
• NIH BEST (Broadening Experiences in Scientific Training) program public reports on PhD career outcomes
• "Biomedical Research Workforce Working Group Report" (NIH Advisory Committee to the Director, public)
• AAAS salary surveys (public summary)

Related Skills

• ors-career-navigation-academic-cv — source CV for the resume translation
• ors-career-navigation-negotiation — full salary negotiation framework
• ors-career-navigation-faculty-interview — if you are running academic and industry searches in parallel
• ors-tailored-resume-generator — automated tailoring of a resume to a job description
• ors-mentorship-teaching-ors-mentorship-goal-setting — IDPs that include industry career goals

Changelog

• 1.0.0 (2026-06-10): Initial adaptation by Pradyumna Jayaram. Compiled from public AAAS, iBiology, Nature Careers, BLS, and Levels.fyi data; cross-referenced to negotiation and resume skills.

Cross-references

Other skills in this category:

• academic-cv
• faculty-interview
• fellowship-application
• negotiation
• tenure-dossier