Novartis is one of the largest pharma companies in the world, a Swiss giant with roughly 75,000 employees and around $50 billion in annual revenue. Over the past several years it has poured money into data and digital: a multi-year AI partnership with Microsoft, an internal research data platform (data42) that pools decades of clinical trial data, and “Biome” digital innovation hubs. Most tech hiring happens through its corporate centers in Hyderabad, Dublin, Prague, Ljubljana, Mexico City, and Kuala Lumpur, plus US sites in Cambridge, MA (research) and East Hanover, NJ (commercial). That context shapes the interviews: fewer LeetCode-hard questions than at a tech company, far more about clinical, regulatory, and commercial data, plus real culture screening around their “unbossed” values.
Process
The loop is fairly standard for a large enterprise, usually three to five stages. It starts with a 20 to 30 minute recruiter screen covering your background, motivation, visa status, and salary expectations. For some early-career and high-volume roles, candidates report a one-way recorded video interview (HireVue-style) before speaking to a human; this does not appear to be universal, so treat it as possible rather than guaranteed. Next come one or two technical rounds, typically with a senior engineer or data scientist on the team. Glassdoor reports commonly describe a resume-driven first technical round (walking through your projects in detail) and a second round of applied questions: live SQL, Python exercises, or discussion of how you would use specific algorithms in a real scenario. After that comes a conversation with the hiring manager, sometimes run as a small panel, which mixes role-specific depth with strategic and collaboration questions. The final stage is an HR or values round that maps your stories to Novartis’s four cultural values (inspired, curious, unbossed, integrity) and covers comp and logistics. Glassdoor pegs the average process at about 35 days, and candidates on Blind and Reddit describe stretches of silence between stages, sometimes pushing the whole thing past two months. Nudging your recruiter is normal and expected.
What they actually ask
- Practical Python, easy to medium. Reported questions include replacing None values in a list, finding the key of the nth-highest value in a dictionary, and pandas data manipulation. Think working-programmer fluency, less hard algorithmic puzzles, though a warm-up on coding patterns still pays off.
- SQL you would actually write at work. Joins, aggregations, window functions, and percentage-of-group calculations show up repeatedly in candidate reports, especially for data engineer and analyst-track roles in Hyderabad.
- Applied statistics and ML for data scientists. Hypothesis testing, regression assumptions, bias and confounding in observational data, and “how would this algorithm work in a real scenario” questions. Expect to explain a model to a non-technical stakeholder, since your day job involves clinicians and commercial leads.
- Pipelines and platforms for data engineers. ETL design, data modeling and warehousing concepts, Spark, and cloud services. Novartis leans on the Azure ecosystem via its Microsoft partnership, with Databricks and Snowflake experience frequently listed in job postings; senior candidates should be ready for a light system design discussion of a batch or streaming pipeline.
- Case-style questions on clinical or commercial data. For example: how would you forecast trial enrollment, structure physician-level sales data, or evaluate a model that flags patients likely to discontinue therapy. Interviewers care about your handling of small, biased, privacy-constrained datasets, not just the math.
- Behavioral questions mapped to Values and Behaviors. Structured, competency-style prompts: a time you disagreed with a decision, worked across functions, or took ownership without being told. Prepare STAR stories in advance; the behavioral guide covers the format.
Levels and comp (2026)
Novartis uses a pharma-style ladder rather than tech levels: roughly Analyst and Senior Analyst, then Data Scientist or Engineer, Senior, Principal, Associate Director, and Director. Director-adjacent titles arrive earlier in your career than they would at a tech company, so do not map them one-to-one to Big Tech seniority. Comp is base plus bonus, with restricted share awards appearing mostly at Associate Director and above; there is no meaningful equity story for most individual contributors. Numbers are sparse and self-reported, so treat them as rough: levels.fyi shows a US software engineer median around $125K total comp (topping out near $200K), and US data scientist medians reported between roughly $150K (Boston area) and $173K nationally, with senior and principal DS roles reaching the low $200Ks. Glassdoor figures land in similar ranges. Geography matters a lot: Hyderabad packages are in local pharma-IT bands (commonly reported in the mid-teens to 40s of lakhs INR depending on level), and Dublin and Barcelona sit on European scales well below US Cambridge or East Hanover. Negotiation room exists on base and sign-on but bands are firmer than in tech.
Prep priorities
Put most of your time into practical SQL and Python, since those are the rounds that actually filter people: window functions, group-wise percentages, list and dictionary manipulation, pandas. A short pass through standard coding interview material at the easy-to-medium level is enough; grinding hard problems is wasted effort here. Second, rehearse your resume projects in depth, including why you made each technical choice and what the business impact was, because at least one full round is built on them. Data scientists should refresh hypothesis testing, regression diagnostics, and one clean story about deploying a model with real users. Data engineers should be able to whiteboard an end-to-end pipeline on Azure or Databricks and defend the design. Finally, write out four to six STAR stories keyed to inspired, curious, unbossed, and integrity, and read up on the company’s AI push (the Microsoft partnership, data42) so your “why Novartis” answer sounds specific rather than generic.
Frequently Asked Questions
Does Novartis ask LeetCode-style questions? Rarely at the hard level. Candidate reports describe practical Python and SQL exercises, easy to medium difficulty, plus discussion of your past projects. Senior engineering roles may add a light design conversation, but there is no whiteboard-algorithms gauntlet.
How long does the process take? Glassdoor’s average is about 35 days, and four to eight weeks is a reasonable expectation. Silence between stages is common; following up with your recruiter after a week or two is fine and often necessary, especially if you hold competing offers.
Do I need a pharma or biology background? No. Most tech hires come from other industries. What helps is showing you can work with regulated, imperfect data and communicate with non-technical scientists and commercial teams. Learning basics like what a clinical trial phase is costs you an evening and reads well in the case round.
How does pay compare to Big Tech? Below it, mostly because equity is minimal. A US senior data scientist might see $150K to $200K total at Novartis versus $300K-plus at a top tech firm. In exchange you get stability, reasonable hours, and strong benefits, and hub-city pay is competitive within each local market.
Similar company guides
Prepping for Novartis? Put it to work:
