ABHAY SHUKLA
Work: Tower C, Diamond District, Domlur, Bengaluru, 560008
(m) + 91 8882007001 (e)
[email protected]SUMMARY
Currently, at Airtel X Labs, I am working on customer propensity models, customer intervention models, and fraud
detection. At Swiggy, I worked on single serve meal recommendation, multi-stakeholder recommendation and order type
classification for pickup & drop service. In R&D team at [24]7.ai, I worked on personalization and intent models with
focus on incrementality and customer experience for clients in Hospitality and Telecom. In RSG Media, I worked on
building Promotion Optimization framework from scratch for major television networks in US.
With experience, I have developed the ability to balance between research and practical mindset to ship models to
production. I am comfortable in working as an individual contributor as well as in teams. I have also successfully
mentored teammates in transitioning to Data Science roles within organizations.
PROFESSIONAL EXPERIENCE
Lead Data Scientist – Airtel X Labs (Sep’20 – current)
Propensity Modelling
Built intent classification model to identify customers with high propensity to buy a new 4G handset
Recommendations by model gave 3x higher conversions compared to spray and pray approach
Unauthorized Document Detection
Built a framework to identify unauthorized documents from recaptured images during customer acquisition
Used semi-supervised approach to tag documents with human-in-the-loop approach to verify tagged data
V1: Trained VGG16 model to classify recaptured documents as un/authorized with ~100% Precision@2k
Customer Intervention Model
Building uplift model to identify customers who need relationship manager’s intervention for predetermined tasks
Senior Data Scientist – Swiggy, Bengaluru (Feb’19 – Aug’20)
Item Recommendations in POP
Automated recommendations pipeline in POP from hardcoded rules to ML driven approach
Conceptualized and implemented a Multi-Armed Bandits based framework for candidate generation for ranking
V1: Built a popularity weighted relevance model for recommendations using item and customer embeddings
V2: Built a pointwise learning to rank model using Gradient Boosting Trees for recommendations
V3: Multi-stakeholder recommendation using Sample-Rank (preprint: arxiv.org/abs/2008.10277)
Representation Learning for Entities
Conceptualized and implemented model for learning Word2Vec based embeddings using only transaction data
Features based on learned embeddings are top performing features in numerous ranking models in production
Order Classification for Uncatalogued Items
Built a model to direct orders with uncatalogued items in Genie to customer care for verification and assignment
Used custom Word2Vec to label data for training and fine-tuned pretrained NLP models in TF Hub for classification
Data Scientist — [24]7.ai, Bengaluru (Jan’16 – Feb’19)
Personalization
Drive incremental conversions for clients by personalized treatment of on-domain visitors
Mining web logs using PIG scripts to extract useful features at page/session level for downstream processing
Building propensity score models to determine whom to target and personalize their experience based on intent
A/B testing
Monitoring of deployed model performance over time and against offline performance achieved during training
Uplift Modelling with Predictive Personas
Targeted treatment of visitors who are likely to convert given chat thus avoiding self-serve cannibalization
Offer treatment on the website based on uplift models for new and repeat visitors
Determine optimal thresholds for models which maximize recall under given offer rate
Model performance monitoring using Kibana and comparing online vs offline results for same time period
ABHAY SHUKLA
Work: Tower C, Diamond District, Domlur, Bengaluru, 560008
(m) + 91 8882007001 (e)
[email protected]Representation Learning for Visitor Journey
Unsupervised feature extraction from web sessions
Built dataset of over 10 million web sessions to train session representation models
Used learned features to train Propensity to Purchase model and benchmarked performance against deployed models
Senior Statistical Analyst — RSG Media, Gurgaon (Aug’14 – Dec’15)
Promo Optimization
Developed Heuristic Hill Climbing Optimization approach to create audience viewership driven campaigns
Captured audience engagement and behaviour on TV Network by Monte Carlo Simulations
Used regression to model reach function and incremental reach analysis across time slots
Developed optimization model for Strategic Planning of episodic and launch promotions to maximize unique viewers
Conducted post campaign analysis to quantify the effect of campaigns and test hypotheses on conversion of viewers
Optimized campaigns generated 10% yield while meeting the goals set by the Network
Co-founder, Consultant — TAMES Consulting Pvt. Ltd. (Mar’11 – Feb’12)
CIN - U74140MP2011PTC027215, DIN - 05112525
Provided engineering consultancy in reclaimed rubber industry
Financial Analyst — Verity Knowledge Solutions (UBS) (July’10 – Oct’10)
Worked as financial analyst in Global Industrial Group for UK and APEJ Markets
PUBLICATIONS
A. Shukla, J. Sathyanarayana and D. Banerjee, "Sample-Rank: Weak Multi-Objective Recommendations Using Rejection
Sampling," in 2020 International Conference on Data Mining Workshops (ICDMW), Sorrento, Italy, 2020 pp. 683-689.
doi: 10.1109/ICDMW51313.2020.00096 (preprint: arxiv.org/abs/2008.10277)
INTERNSHIPS
Real Time Proxy Net Promoter Score from Unstructured Data — Hewlett Packard, Bangalore (Mar’14 - Jul’14)
Reviewed the WordNet based algorithm developed by Global Analytics and proposed 3 alternate methodologies
Proposed Methodologies:
SVM with unigram and bigram features to classify reviews into promoters, passives and detractors
Rule-based unsupervised classification by scoring reviews using SentiWordNet lexicon
Sentiment Aggregation algorithm to identify class of documents from feature sentiments
Used bootstrapping to compute confidence intervals for NPS estimates
Proposed methodologies resulted in accurate NPS estimates compared to algorithm developed by Global Analytics
EDUCATION
Year Institute, City Degree/Certificate CGPA/%
2012-14 Indian Statistical Institute, Kolkata M.Tech, QROR 83%
2010 IIT, Delhi B.Tech, Production & Industrial Engineering 7.57/10
2005 Aklank Public School, Kota Class XII,CBSE 83%
2003 Central Academy, Rewa Class X, CBSE 84%
EXTRACURRICULAR ACTIVITIES & ACHIEVEMENTS
Awarded Certificate of appreciation for improving the effectiveness of the IIT Delhi NSS campaign in slums
Awarded Certificate of appreciation by NGO Udayan for teaching adopted children for a period of one year
Direction and Story Writing (with renowned animator and photographer Nitin Donde) (Aug’10 - Nov’10)
Led a team of 14 to complete an animation film, featuring original characters, using cut-stop motion technique
Editor and Designer, iitstories (Media Platform for Start-ups), New Delhi
Edited and published articles and interviews related to start-up companies by IIT alumni & students
