Pratyush Agarwal

AI Researcher

Bengaluru, Karnataka, India5 yrs experience

Key Highlights

  • Reduced revenue leakage by $4K/day through automation.
  • Generated $3.7M savings and $10M revenue boost.
  • Designed scalable dashboards saving 2 analyst-days weekly.
Stackforce AI infers this person is a Data Science professional with expertise in E-commerce and Healthcare analytics.

Contact

Skills

Core Skills

A/b TestingExperimental DesignData VisualizationData Pipeline AutomationData AnalysisBusiness InsightsCausal InferenceStatistical ModelingStatistical AnalysisDeep Learning

Other Skills

MySQLSnowflakeSQLHealth EconomicsReal World Evidence (RWE)RPythonTensorFlowVisual StorytellingMicrosoft ExcelMicrosoft Power BIReporting & AnalysisBusiness Data ManagementStakeholder ManagementBusiness Metrics

About

- Result-driven Data & Analytics professional with a proven track record of driving business impact through experimentation, automation, and data-driven decision-making. - Skilled in A/B testing, causal inference, statistical modeling, and real-time analytics, with hands-on experience in product optimization, growth analytics, and supply strategy. Product & Business Analytics - Designed A/B & switchback experiments optimizing product adoption & engagement - Eliminated $4K/day in revenue leakage, reducing restaurant attrition by 15% MoM through automated insights - Built scalable dashboards, removing manual tracking & saving 2 analyst-days per week - Enhanced category conversion (+1pp) and order contribution (+3pp) via consumer funnel optimizations Real-World Evidence and HEOR: - Developed statistical models & real-world evidence (RWE) enabling $3.7M in savings & a $10M revenue boost - Improved model accuracy by 25% and reduced error by 15% through stable balancing weights in R and non-parametric bootstrapping, eliminating baseline bias and outperforming conventional propensity-based methods - Automated data workflows, cutting turnaround time by ~30% and reducing errors to zero. - Published research in Medical Devices: Evidence and Research (CiteScore 3.0) on clinical outcome comparisons. Core Competencies ✔ A/B Testing & Experimentation ✔ SQL, Python, R: Data Wrangling and working with large streams datasets ✔ Data pipeline creation, reporting, and visualization with PowerBi and Tableau ✔ Causal Inference & Statistical Modeling ✔ Advanced Excel Analysis (VBA, Index-Match, Pivot) ✔ Data Pipeline Automation | Real-Time Analytics | Product & Business Growth Analytics ✔ Data Visualization (Power BI, Tableau) | KPI Design | Cross-functional Collaboration & Mentorship

Experience

5 yrs
Total Experience
2 yrs 5 mos
Average Tenure
1 mo
Current Experience

Uber

Data Science

Mar 2026Present · 1 mo · Bengaluru, Karnataka, India · Hybrid

Swiggy

2 roles

Analytics Manager

Oct 2024Mar 2026 · 1 yr 5 mos · Bengaluru, Karnataka, India · Remote

A/B TestingExperimental Design

Senior Data Analyst

May 2023Sep 2024 · 1 yr 4 mos · Bengaluru, Karnataka, India · Remote

  • Designed and executed A/B and switchback experiments, optimizing product/business metrics and informing key initiatives like On-Time Guarantee, Pocket Hero, Swiggy Occasions
  • Automated real-time performance dashboards, eliminating manual tracking and saving 2 analyst-days per week
  • Diagnosed and resolved $4K/day revenue leakage in restaurant billing & offer configurations
  • Reduced restaurant attrition by 15% MoM through weekly automated insights via email
  • Improved category order contribution by 3pp and category conversion by 1pp with 20K iOPD by identifying consumer journey gaps and implementing targeted funnel optimizations
  • Developed a supply-scoring framework, boosting hyperlocal sourcing efficiency and food widget conversions by 0.2pp
  • Established instrumentation for backend and frontend event logging, enabling real-time analytics and data-driven decisions
MySQLSnowflakeA/B TestingData Visualization

Mu sigma inc.

3 roles

Decision Scientist 3

Promoted

Jan 2022Apr 2023 · 1 yr 3 mos

  • US Fortune 50 Healthcare Company | HEOR & RWE:
  • Distinguished causation from association (causal inference) using regression models, covariate balancing (using Stable Balancing Weights, Propensity Score Matching/Weighting/Stratification), cumulative incidence functions, hypothesis testing, and non-parametric bootstrapping to prove equivalence of our product with a comparator
  • Generated Real-World Evidence (RWE) using analytics on patient level Real-World Data (RWD) sources like EMR (Mercy), EHR (Optum, Loopback), claims (IBM, SAF, Optum), and administrative (Premier) to measure HCRU, burden of illness, treatment patterns, effectiveness, and clinical outcomes
  • Supported product commercialization, reimbursements, and value proposition of client’s products for HEMA stakeholders by producing statistical reports and doing extensive EDA on NLP datasets
  • Developed ADaM datasets, TLFs, safety packages, research protocols, clinical study reports with Descriptive and Statistical Inference analysis and proactively worked using vast real-world data sources to generate RWE
  • Designed automated solutions for data extraction and output generation processes through SQL, R, VBA, and Python macros that reduced turnaround time for insight generation from 3-4 days to 1-2 hrs and made deliverables error-free
Health EconomicsReal World Evidence (RWE)Causal InferenceStatistical Modeling

Decision Scientist 2

Jan 2021Dec 2021 · 11 mos

  • Bangalore Police (Cybercrime Division) | Cybercrime Insight Generation:
  • Collaborated with IPS officers from cybercrime division to understand their problem universe, hypothesis, project expectations, existing bottlenecks, and consequently developed iterative mock-ups to establish project scope
  • Leveraged VBA scripts to de-identify data, perform automated cleaning, and generated features using 6 cyber-crime datasets sourced from multiple platforms with varying time lags
  • Visualized KPIs in excel to identify stressed areas in cyber-crime reporting and investigation lifecycle using various charts and context filters

Trainee Decision Scientist

Sep 2020Dec 2020 · 3 mos

  • US Fortune 50 Home Improvement Company | Retail customer experience enhancement:
  • Using customer segmentation and developing key performance indicators, improved customer experience for a Fortune 50 US-based home improvement retailer
  • Identified disparities between queue wait-time for different customer segments where higher LTV customers had to wait longer in queues than lower LTV customers because of their usual cart size and store’s counter design
  • Created Tableau dashboard to visualize identified KPIs like cart value and queue wait-time for different counters to establish association and drive changes in counter design to reduce wait time
  • US Fortune 50 Healthcare Company | HCP detailing during COVID:
  • Stakeholders during the COVID-19 pandemic faced challenges in identifying healthcare professional clinics that remained open for "detailing"
  • Longitudinal multi-year data from 4 different data sources with different latency were used and micro-trends were analyzed to evaluate the odds of a clinic being open using different weighting techniques

National university of singapore

Research Intern

Jan 2019Jan 2019 · 0 mo · North East, Singapore · On-site

  • Worked under Dr. Wee Kek Tan and Dr. Hsiang Hui Lek to develop a facial emotion-based music generator
  • A CNN model which was 15 layers deep was created and trained on AffectNet dataset to classify human emotions into 1 of 5 categories
  • Based on the detected emotion, 4 different pre-trained RNN models (trained on the hand-labelled chorus of 1000 songs) were used to generate music based on recognized emotions
TensorFlowDeep Learning

Education

Kalinga Institute of Industrial Technology, Bhubaneswar

Bachelor of Technology - BTech — Computer Science

Stackforce found 100+ more professionals with A/b Testing & Experimental Design

Explore similar profiles based on matching skills and experience