Oct 2024 – Jun 2025 · 8 mos · Seattle, Washington, United States · Remote

• Data Scientist | Azure Cost & BI Platform (https://azure-insight-collective.lovable.app/)
• Engineered a real-time anomaly detection pipeline using PySpark, Databricks, and Kafka to monitor cloud resource KPIs (e.g., cost/hour, CPU usage), flagging underutilized or misconfigured resources early, preventing sudden cost spikes.
• Conducted statistical analysis and thematic clustering on a 67-participant user survey to quantify key pain points (e.g., tool fragmentation), shaping prioritization for a platform redesign, and improving user experience.
• Built a unified Business Intelligence dashboard aggregating cost, savings, and sustainability KPIs, which improved insight-to-action time by 40% and CSAT by 35% among pilot users.

Jul 2024 – Aug 2024 · 1 mo · Seattle, Washington, United States · Remote

• Data Scientist | Consumer Insights, CPG
• Applied K-Means clustering to psychographic and usage data of 4000+ Gen Z survey responses to segment customers into personas, enabling personalized speaker campaigns via social media, projected to boost customer engagement by 20%.
• Developed logistic regression model (AUC = 0.81) and ran hypothesis tests on survey responses to identify loyalty drivers; found that higher sound and battery ratings increased repurchase odds by 20-25%, informing Beats’ customer retention strategy.
• Presented insights to marketing leadership using interactive Tableau Dashboards, visualizing High Lifetime-Value (LTV) customers and higher loyalty among premium buyers (22%) to scale revenue by focusing on these user segments.

Jun 2024 – Aug 2024 · 2 mos · Seattle, Washington, United States · Hybrid

• Developed and deployed AI-driven quiz tool using Gemini LLMs with Google Cloud, Streamlit, Vertex AI, and LangChain to generate dynamic quizzes with real-time feedback, improving user learning outcomes by 30%.
• Devised a data ingestion pipeline with ChromaDB and PyPDFLoader to automate PDF processing and text chunking, boosting data transformation accuracy by 15%.
• Built a RAG-based chatbot by integrating LLMs with Streamlit and engineering custom NLP pipelines, increasing user satisfaction by 15% through personalized, contextual responses (Prompt Engineering).

Mar 2023 – Sep 2023 · 6 mos · Mumbai Metropolitan Region · On-site

• Data Scientist | EdTech
• Analyzed interview data to identify skill gaps in undergraduate students and collaborated cross-functionally with Technical, and Marketing teams to launch new courses (Soft Skills, Resume Building), driving a 20% increase in revenue in 6 months.
• Built a predictive-modeling recommender using XGBoost on post-launch student data to predict course completion likelihood which led to personalized course paths for each student, boosting employability by 17%.
• Designed and executed an A/B test to validate hypothesis that adding social proof (e.g., star ratings) to course cards would boost engagement; confirmed a statistically significant 12% lift in 'Enroll' Clickthrough Rates, leading to a site-wide rollout.
• Reduced a critical student-facing page load time by 60% (12s to 4s) by optimizing SQL queries and normalizing database schema of 300k+ student data; decreased related support tickets by 70% during peak periods.

Jan 2023 – Mar 2023 · 2 mos · Mumbai, Maharashtra, India · Hybrid

• Constructed an Object detection model leveraging YoloV5 architecture for parking lot management system enhancing portal’s overall capabilities by adding 3 new functionalities.
• Performed data scraping for data collection, acquiring extensive datasets of 37 GB for said project containing pictures of parking lots across all weather conditions to better train the model
• Orchestrated the model training process, tweaking hyperparameters and ensuring the incorporation of data augmentation techniques, such as blur, contrast and brightness modification, and rotation, to attain optimal accuracy of 92%.

Jun 2022 – Nov 2022 · 5 mos · Mumbai, Maharashtra, India

• Performed feature engineering on data from a portfolio of 15+ cloud kitchens (order volume, delivery times, food costs) to train a random forest model (R² = 0.67) that predicted a composite performance score and identified 3 “happy-but-underperforming” kitchens; prompted a strategic pivot to optimize their operational workflows, increasing order capacity by 25%.
• Led a 4-intern team to build an NLP pipeline that classifies 20K+ kitchen reviews; the fine-tuned BERTweet model flagged high-risk feedback with 81% accuracy (F1: 0.79), cutting manual review effort by 40%.
• Developed a SQL-based data validation and anomaly detection pipeline to clean and normalize 50k+ kitchen inventory logs; flagged input errors and abnormal stock activity, reducing stock discrepancies by 19% across 3 kitchens.

May 2022 – May 2023 · 1 yr · Mumbai, Maharashtra, India · On-site

• Administered InIT Hackathon 2023 coordinating with a team of 50 members.
• Mentored 25 juniors on machine learning projects by conducting lectures on trending topics.

Feb 2022 – May 2022 · 3 mos · Mumbai, Maharashtra, India · Hybrid

• Centralized fragmented departmental data into a Microsoft SQL Server backend via Power Query ETL pipelines and Excel (XLOOKUP, Pivot Tables), streamlining monthly reporting and reducing manual processing by 3+ hours.
• Drove a 15% increase in sales by identifying high-margin product bundles through market basket analysis on 2 years of transactional data and aligning their promotion with peak seasonal demand using time series forecasting (ARIMA).

Mar 2021 – May 2023 · 2 yrs 2 mos · Mumbai, Maharashtra, India

• Designed, procured and managed the battery pack, battery management system and motor controller of the kart.
• Collected data to tweak gear ratio for optimizing track speed.
• Using Arduino Uno and magnetic reed switches, displayed kart velocity, battery percentage, and laps remaining.
• Drafted comprehensive cost and design reports to support innovative enhancements.