CV HimanshuJain ML Engineer

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

London, United Kingdom

+44 7880991472
Himanshu Jain Himanshuj2018@email.iimcal.ac.in
Senior Machine Learning Engineer linkedin.com/in/himanshu-jain-7382869b

Over 9 years, I've led high-impact ML projects in financial services, building and deploying scalable models (1+ billion data points). My expertise spans the entire
ML lifecycle, ensuring successful deployments from experimentation to production. Led the successful adoption of Dask for distributed computing and Sage
Maker pipelines, resulting in 90% reduction in processing time.

Education
Master of Science in Financial PAVE Fintech| Open Banking | Sr. ML Engineer (Nov 2022 – Present)
Engineering (2024 – 2025) (Online)
World Quant University (WQU) Transaction Enrichment- Torch based NLP model categorizes a staggering 1Bn+ financial transactions.
• Achieved remarkable results with a multi-class, 2-stage neural-network based classifier, boasting over 97%
Master’s in Data Science & Business accuracy and < 0.08 entropy loss, effectively predicting 250+ income and expense categories.
Analytics (2016 – 2018) (CGPA 8.9/10) • Elevated dataset coverage by 70% by spearheading data preparation efforts, using multi-layer perceptron
Globally Ranked 14 (QS World) classifier, web searches and CHATGPT prompt-engineering.
Indian Institute of Technology (IIT), • Developed custom LLM model by finetuning Mistral-7B using QLoRA for named entity recognition of
Kharagpur transactions’ data with a F1-score of 86%.
Indian Institute of Management (IIM), • Achieved a remarkable 90% reduction in pipeline execution time by adopting DASK for distributed
Calcutta
computing and GPUs for faster computation.
Indian Statistical Institute (ISI), Kolkata
• Leveraged multi-stage Docker builds for compact image size across 10+ applications, employing dedicated
images for Python dependencies, source code, and pipeline execution.
B.E (2009 – 2013) (CGPA 8.2/10)
NSIT, Delhi University, India Open Banking Pipeline
• Architected a low-latency and scalable microservices using AWS Lambda to generate 1300+ credit
affordability features for B2B partners, enabling faster decision-making.
Achievements
• Orchestrated multiple services into serverless workflows using AWS Step functions for seamless batch
• Ranked 19 worldwide in Data Science processing of banking transactions for over 300,000 users.
Game, Paris among 220 teams hosted by • Orchestrated data integration using AWS Glue to sync 3rd-party partners data from RDS to s3 data lakes.
Kaggle and Microsoft
• Top 0.1 percentile in Graduate Aptitude JP Morgan Chase| Asset Management | Sr. ML Engineer (Dec 2020 – Nov 2022)
Test among 2,80,000+ entrants
Knowledge Graph – Led the greenfield project of building the graph for Wealth Management Clients
• Built Entity Resolution service by leveraging alternative dataset like Wealth-X, FactSet, and Crunchbase,
Skills expanding customer information coverage for analysts by a staggering 300%.
• Languages: Python, PySpark, SQL • Designed a connection strength model deriving insights from Professional, philanthropic, and hobbies data.
• Enabled efficient data representation by deriving graph embeddings using FastRP (Fast Random
• ML Skills: NLP, Deep Learning (Mask R-
Projections) algorithm in Neo4j.
CNN, Transformers), Computer Vision
(OpenCV), Recommendation Systems, Email Intent– Pioneered an ML-First email analysis product, modernizing the usage of 60+ shared mailboxes.
Pandas, NumPy • Designed an Email Intent classifier with over 98% accuracy, effectively classifying non-actionable and
• Automation: Airflow, Docker, Dask actionable emails to reduce operation time over 90%.
• Cloud: AWS Elastic Container, Sage • Developed a cutting-edge Text Sentiment model utilizing in-house Fin-BERT and dictionary-based models.
Maker pipelines, Glue, Lambda, Step
functions Zilingo|E-commerce Fashion Start-up | Sr. Data Scientist (Jul 2019 - Dec 2020)
• CI/CD Skills: GitHub Actions, Pre-
commit, Unit & Integration testing Fashion Recommendation- Content + Collaborating filtering based product recommendations across
• Code Versioning: AWS Elastic Container all South-east Asia regions
Registry, Bitbucket, Git • Used ResNet50 CNN architecture to extract the product features for product Catalogue of 1Mn+ products.
• Databases: Neo4j, PostgreSQL, NoSQL • Retrieved similar items by using Approximate Nearest Neighbors making it 100x faster than older methods.
• Successfully conducted AB testing, resulting in a solid 7% Click-Through Rate (CTR)
• Data Visualization: QlikView, Tableau
B2B Merchants Risk Profiling: Credit scoring based on merchant’s transactions & demographics.
• Achieved 63% Precision and 62% Recall on new Credit loan applications for Q4 2019-20.

Barclays | Data Scientist (Oct 2017 -Jul 2019)

Digitizing PPI Documents – Built an Optical Character Recognition (OCR) engine to digitize 24 Mn documents span over 20 years.
• Achieved 80% estimated Accuracy on scanned applications and estimated 2Mn Pounds cost saving to the business
Smart Consumer Lending – Built Loan Propensity model to target customers based on demographics and spending data.

KPMG| Analytics Consultant (Aug 2013 – Jun 2016)

• Single point of contact to deliver Logistics, Production & Financial KPIs through Dashboards via 4-tier Architecture

You might also like