Razorpay's Machine Learning hiring loop is a production engineering gauntlet disguised as a model-building role. Their acceptance rate for external L4-L5 ML Engineer candidates is sub-2%, with the committee auto-rejecting any resume lacking direct proof of owning a live model's end-to-end lifecycle—from feature pipeline SLA to business PnL impact. They filter for engineers who speak in Razorpay's internal KPIs: Payment Success Rate (PSR), Checkout Drop-off %, Merchant Fraud Loss Rate (MFLR), and Automated Reconciliation Efficiency. Your resume must pass Keka HR's keyword scoring, then survive a 90-second hiring manager scan calibrated against their internal ‘Impact vs. Scope’ rubric. Generic ML project bullet points are noise.
ATS Insider Intelligence
How Keka HR Actually WorksKeka HR at Razorpay parses PDFs optimally but strips all styling, rendering multi-column layouts as garbled text. It heavily weights keyword frequency in the most recent 24 months of work history, not the Skills section. Recruiters pre-filter using knockout tags: ‘production ML,’ ‘A/B testing,’ and ‘online serving’ are mandatory for ML Engineer roles. The system generates a recruiter scorecard with auto-populated fields for ‘Tech Stack Match’ and ‘Relevant Tenure.’ Crucially, Keka flags resumes with keyword stuffing (e.g., repeating ‘machine learning’ 10+ times) as low-quality, pushing them down the ranking. Hyperlinks to GitHub or portfolios are often stripped before committee review.
🎯 ATS Keyword Arsenal
Razorpay • Machine Learning Engineer • Keka HR — Click any keyword to copy it
⚡ Technical Skills
🔧 Tools & Platforms
🧠 Behavioral / Soft Skills
🏢 Domain Expertise
See how many you're already using 👇
Applying without auditing your resume against Razorpay's Keka HR keyword rubric is like shipping a model without validation—it fails silently. Your resume gets an auto-generated score; a few missing terms can place you below the recruiter's view threshold.
Test Your Resume Against Razorpay's Machine Learning Engineer Keywords
Paste your resume bullets below to see how many of Razorpay's 39 target ATS keywords and impact metrics you match right now.
Expert Resume Tips for Razorpay
Front-load every bullet with the business KPI you moved, not the model architecture.
Razorpay's L5 rubric requires 'demonstrated impact on key business goals.' Start with: 'Increased Payment Success Rate by 2.1%...' or 'Reduced Merchant Fraud Loss Rate by $1.2M annually...' Then detail the technical method. This forces alignment with their product-focused ML culture and gets past the hiring manager's impact filter.
Why this matters at Razorpay
Hiring committees score 'Business Impact' as 40% of the total candidate score. A KPI-first structure proves you think in terms of Razorpay's north-star metrics, not academic novelty.
Explicitly name Razorpay-equivalent system components like 'Online Feature Store' and 'Real-time Inference Service.'
Do not say 'built a feature pipeline.' Say 'built an Online Feature Store using Redis and DynamoDB to serve 10K+ low-latency features for real-time fraud scoring.' This demonstrates you understand the production stack required to serve models at Razorpay's scale (10M+ daily transactions).
Why this matters at Razorpay
This signals you can ramp up without hand-holding. Their ML platform team uses these exact architectural patterns; using the correct terminology triggers a 'stack familiarity' tag in the recruiter's Keka scorecard.
Quantify latency in milliseconds and throughput in QPS for any inference system work.
Razorpay's payment stack has strict SLOs. A bullet like 'Optimized fraud model inference latency from 120ms to 75ms P95, supporting 5K QPS' is high-signal. Generic 'improved performance' gets ignored. Specify if you used model quantization, pruning, or a specific serving framework (TorchServe, Triton).
Why this matters at Razorpay
The SDE/ML track explicitly evaluates 'system design for scale.' Latency and QPS numbers prove you've operated under constraints similar to Razorpay's high-throughput, low-latency payment environment.
Describe A/B test results using their internal metric names and confidence intervals.
Write: 'A/B tested a new checkout flow recommender, achieving a 1.8% lift in Checkout Conversion Rate (p < 0.01, 95% CI) over 4 weeks.' Avoid vague 'improved user experience.' This shows rigor and familiarity with their product experimentation culture.
Why this matters at Razorpay
Razorpay's growth and product teams mandate statistically sound experiment reporting. This phrasing indicates you can collaborate directly with them and won't ship models without measurable validation.
Place 'Machine Learning Engineer' as your exact job title, even if your official title was different.
Keka HR's initial filter matches on role title keywords. If your title was 'Data Scientist' or 'Research Engineer,' add a parenthetical: 'Data Scientist (Machine Learning Engineer).' This ensures the system doesn't reject you based on a semantic mismatch before a human sees your production ML experience.
Why this matters at Razorpay
Recruiters often search by the exact role title in Keka. A mismatch can bury your profile. This is a known workaround used by internal referrals to ensure proper categorization.
Before vs After: Real Bullet Rewrites
These are the exact bullets that get filtered vs. the ones that get through Keka HR and land interviews.
⚡ Insider Counter-Intuition
Razorpay's hiring bar for ML Engineers is 80% software engineering and 20% modeling. A candidate who built a simple logistic regression model but engineered a bulletproof, low-latency, monitored serving system will rank higher than someone with a state-of-the-art model deployed via a fragile, hand-maintained script. Your resume must scream 'Software Engineer who specializes in ML systems,' not 'Researcher who can code.' This is why they prioritize Kubernetes, CI/CD, and SLOs over novel algorithms.
Mistakes That Get Machine Learning Engineers Rejected at Razorpay
FAQ: Machine Learning Engineer at Razorpay
Does Razorpay's Keka HR ATS reject resumes for having hyperlinks?
Yes, hyperlinks are often stripped during parsing and are invisible to the hiring committee. Never rely on 'see my portfolio at...' or a GitHub link to convey critical information. Assume only the plain text of your PDF will be reviewed. Embed your most impressive project metrics directly in the bullet points.
What's the difference between a Data Scientist and Machine Learning Engineer resume for Razorpay?
Razorpay hires ML Engineers for production system ownership and Data Scientists for insights and experimentation. Your ML Engineer resume must emphasize: online serving, latency, scalability, CI/CD, MLOps (MLflow, Airflow), and collaboration with platform/SRE teams. A Data Scientist resume would focus more on SQL, dashboards, and statistical analysis. Confusing the two leads to a role-fit rejection.
How important are publications or Kaggle rankings for a Razorpay ML Engineer role?
Largely irrelevant for L4-L5. Razorpay's ML Engineering track values production deployment and business impact far more than academic accolades. A Kaggle Grandmaster with no production experience will be rejected. If you have them, list them briefly at the bottom. Your 3-4 top bullets must be about work in a live, user-facing system.
Should I mention experience with UPI or other Indian payment systems?
It's a massive domain-signal advantage. Mentioning specific experience with UPI's flow, NPCI regulations, or the Indian fintech stack (Aadhaar, PAN, etc.) immediately demonstrates you understand the core context of Razorpay's business and can contribute faster. Weave it into domain-specific keywords or a short 'Domain Knowledge' line.
How does Razorpay's hiring committee level candidates (L4, L5, L6) based on a resume?
They use an internal rubric: L4 (executes projects), L5 (owns complex projects end-to-end and influences roadmap), L6 (sets strategy for multiple teams). Your resume must show scope: L5 requires multi-model ownership, cross-team collaboration, and clear PnL impact. L6 requires shaping the technical vision for a whole domain like Risk or Growth. Tailor your achievement scale accordingly.
Related Resume Guides
Other Roles at Razorpay
Check Your Razorpay ATS Score
See exactly how many of these 39 keywords you're using right now.
Run ATS Check Free Build Resume with AI🔍 ATS Being Used
Keka HR
Optimize specifically for Keka HR to beat the automated filter before a human even sees your resume.
Stop Guessing. Start Matching.
Upload your resume and see exactly how many of these 39 Razorpay-specific keywords you're already matching — and which gaps are costing you the interview.