Ashish Dhiman

Optimisation, Explainable AI, Data Science, Credit Risk, Geopolitics Enthusiast

📧Email / 🔗LinkedIn / 🔗GitHub / 📞+1-4045090254

♾️ Objective

To augment my Data Science knowledge base and pursue expertise in Optimisation Methods and Explainable Machine Learning. I aspire to develop my skill-set and networking connections that will act as a stepping stone in my career and help open new pathways for me, like Applied Research roles in coveted firms or postgraduate education. These skills will also help me contribute to the availability of digital mental health services.

👩🏼‍🎓 Education

Georgia Institue of Technology, Atlanta (Aug 2023 - Dec 2023)
MS in Analytics: Computational Data Science and Analytical Modelling Track | CGPA: 1st semester

Indian Institue of Technology, Kharagpur (Jul 2015 - May 2019)
B.Tech(Honours) in Aerospace Engineering, with Micro Specialisation in Optimisation Theory | CGPA 8.32/10 | Class Rank 2

👩🏼‍💻 Work Experience

Assistant Manager @ American Express India, Gurgoan, India (Aug 2021 — Present)
Consumer Collection Deciscion Science, Credit and Fraud Risk

• Envisaged and implemented a novel domain of information source by transforming the trade level supplementary and joint account linkages present on the Experian bureau’s end to create an inter-customer linkage relationship graph.
• Leveraged the inter-customer relationship graph to create modeling attributes based on primary and secondary linkages, helping prevent annual credit default of $2.5 million, along with better identification of Credit Bust Out(CBO) fraud.
• Implemented and Investigated the appropriateness of two feature selection algorithms, namely Gradient Boosted Feature Selection(GBFS) and min Redundancy and Maximum Relevance(mRMR), for application in Credit Risk context.
• Part of the team that implemented an automated tool to parse metadata like name, contact, education, skills, etc., from the resume pdf files using a combination of regex rules, zero-shot classifier and Named Entity recognition (NER) model
• Technologies used: Python, Spark, SQL, AWS, GBDT, SAS, NER.
  
  

Risk Analyst @ American Express India, Gurgoan, India (Jul 2019 — Jul 2021)
Consumer Collection Deciscion Science, Credit and Fraud Risk

• Implemented a Payment Relief model, on Experian’s end, to identify customers enrolled in Covid-19 Payment Relief programs of external Creditors, by using tradeline level history of monthly balances, payments and account status codes.
• Developed an automated pipeline to, parse risk triggers from the Bureau related to External Covid Relief Programs, generate an MIS report, and share it with the leaders in real time, with the anomalous sections highlighted.
• Improved the accuracy of dollar predictions of External Payments model by 7%, through implementing Synthetic Minority Over sampling technique(SMOTE), for people with actual payment reports available.
• Technologies used: Python, Spark, SQL, AWS, GBDT, SAS.
  
  

📖 Research Experience

Research Intern @ University of Otago, Dunedin, New Zealand (May 2019 — Jul 2019)
Professor Stephen Cranefield, Department of Information Science

• Researched on a novel approach to learn arbitrarily complex Norms from a Normative Language, expressed as a Probabilistic Context-Free Grammar(PCFG), in the setting of Multi-Agent Societies. Paper published in proceedings of IJCAI’21.
• Modeled Norms as expressions from a recursive grammar space (PCFG) and applied Monte Carlo Markov Chain (MCMC) technique to sample the posterior distribution of Candidate Norms, by observing a task under a True Norm.
• Formulated an approach to analyse the convergence of MCMC chains, by transposing the expressions from the grammar to instances from tree kernels, which allowed for the calculation Gelman Rubin Rˆ convergence statistic measure.
• Technologies used: Python, Bayesian Learning, MCMC, Tree Kernels.
  
  

Bachelor’s Dissertation @ IIT Kharagpur, West Bengal, India (Aug 2018 — May 2019)
Professor Sujoy Bhattacharya, Vinod Gupta School of Management, IIT Kharagpur

• Examined Enhanced Index Tracking (EIT), a class of Optimal Portfolio Selection problems, and modified the heuristic Kernel Search framework to improve the return per unit risk performance by 12% over the vanilla Kernel Search method.
• Formulated the EIT as a Mono-Objective Linear optimisation, leveraging a weighted Risk & Return objective function.
• Used the dimension-reduction strategies of Nonnegative Principal Component Analysis (NPCA) and Nonnegative Matrix factorisation (NMF), to emphasize the long-run performance of securities while suppressing the short-term noise.
• Technologies used: Python, R, Linear & Convex Optimisation, NMF/NPCA, Kernel Search Methods.
  
  

⚙️ Internships

Risk Management Intern @ American Express India, Gurgoan, India (May 2018 — Jul 2018)
Consumer Collection Deciscion Science, Credit and Fraud Risk

• Awarded Pre-Placement Offer (PPO), for exceptional performance shown during internship.
• Improved performance of External Contact propensity model, by 16%, along with a reduction in variables used by 22%.
• Implemented an automated process in Python and Shell, for optimizing the parameter tuning step of Gradient Boosted Decision Trees (GBDT) model, by leveraging Bayesian Optimisation and Grid Search methodology.
• Technologies used: Python, SQL, GBDT, SAS, Tableau.
  
  

Decision Science Intern @ Quantiphi Analytics Pvt. Ltd., Mumbai, India (May 2017 — Jul 2017)
Athena’s Owl

• Adjudged Eklavya: Best Intern-2017, for exceptional work on Object detection (R-CNN) workstream of Athena’s Owl.
• Implemented an all-encompassing clustering module using Python & Tableau, reducing the time from 3 weeks to 4 days.
• Developed a module to scrape meta-tags of TV Shows, and infer critical episodes from it, using Selenium and Beautiful Soup. The meta tagging product Athena’s Owl received widespread acclaim at Google Cloud Summit-2017.
• Technologies used: Python, R, Tableau, OpenCV.
  
  

🥇 Competitions

Ranking Mutual Fund Houses in India @ Inter Hall Data Analytics (IIT Kharagpur- 2018)

• Captained a gold winning team of 20 members, for developing a comprehensive framework to rank Mutual Fund Houses.
• Applied LSTM for multivariate temporal forecasting of Net Asset Values (NAV) of mutual funds, and Vector Auto Regression (VAR) for quantifying responsiveness of the fund houses against macroeconomic anomalies and shocks.
• Technologies used: Python, LSTM, VAR, Clustering, Tableau.
  
  

Ranking ATM providers in California @ Inter Hall Data Analytics (IIT Kharagpur- 2017)

• Developed Gold-Winning model to evaluate Competitive Strength of ATM providers, with ATM locale & demographics.
• Designed Market Share metric as the measure of Competitive Advantage for each zip-code, calculated using population demographics of zip-code, and proximity of ATMs to economic hubs, information scraped with Google Maps API.
• Technologies used: Python, GMaps API, R, Clustering, Tableau.
  
  

🖥️ Computer Skills

• Programming languages: C / C++ / Python / MATLAB / Scala / SQL / R / SAS / Octave / LATEX / Bash
• Software/Frameworks: AWS / Spark / Hive / Tableau / Deepnote / Google Colab

📜 Publications & Certifications

1. Cranefield, S. & Dhiman, A. Identifying Norms from Observation Using MCMC Sampling (ed Zhou, Z.-H.) Main Track, 118–124. https://doi.org/10.24963/ijcai.2021/17 (Aug. 2021).
2. Machine Learning 2016. https://coursera.org/share/50963ab3926e11775951a32b5cd82d28.
3. Fundamentals of Quantitative Modeling 2017. https://coursera.org/share/587a2a6d86792ad9de9d88a269cc0297.
4. Basic Data Descriptors and Statistical Distributions 2018. https://coursera.org/share/4f972c591b3fb10cb13a0a18f16f679f.
5. Bayesian Methods for Machine Learning 2020. https://coursera.org/share/8b85193f4feff4b50a434bd0ddba3f00.
6. Machine Learning with Python 2020. https://coursera.org/share/4b8ee67654d6fc864b75a0771a0e48b0.
7. Big Data Analysis with Scala and Spark 2020. https://coursera.org/share/d1afd1c66d70f68b0ddb9731aa580a34<.

🏆 Accomplishments

• Achieved an All India Rank within Top 0.5% in the JEE 2015, among the 1.3 million registered applicants.
• Received letter of appreciation (2015) from HRD Minister, Government of India, for exceptional performance in Class 12.
• Awarded Gymkhana Award, 2019 for outstanding contribution to Technology events, during the 4 years stay at IIT KGP.
• Awarded Analyst of Quarter (Q3’2020), and Senior Vice President Award (Q1’2021) by American Express.
  
  

📌 On The Side

• Conducted biweekly education sessions in the neighbouring rural areas of IIT-KGP as an NSS Volunteer b/w 2015-2017.
• Guided 5 juniors on the emotional & academic fronts in their first year at IIT-KGP as a Student Welfare Group volunteer.
• Inter Hall events (IIT Kharagpur): 🥇Gold in Ad-Design 2017, 🥇Gold and 🥉Bronze in Case Study 2018 & 2017 respectively.
• Secured 3rd prize, HULT(2018) campus round,for a product to increase efficiency of last-leg vaccination outreach in India.
• Designed a portable & durable water filter, suitable to be used in rural parts of India for access to clean drinking water.
• Selected in final 10 teams of Economist, a pan India case study competition on analysis of Reliance Jio (Kshitij, IIT KGP).