Machine Learning Data Scientist

Shravan Kumar Venkannagari

ć *******.****@*****.*** Ħ +1-425-***-**** ] linkedin.com/in/shra1v Summary

Experienced Data Scientist and AIML Engineer with a proven track record of developing and implementing innovative solutions to complex problems across industries. Expertise in machine learning, deep learning, computer vision, Natu- ral Language Processing, Generative AI (GenAI), Large Language Models (LLMs), and Retrieval-Augmented Generation

(RAG) pipelines. Proven ability to deliver significant business impact through data-driven insights and predictive modeling. Skilled in translating complex technical concepts into actionable solutions. Skills

Programming Languages: Python • C++ • C • SQL • Matlab Frameworks and Libraries: AWS• Azure • AKS • Google Cloud • GCP • GKE • OpenShift • Docker • Kubernetes • Ten- sorFow • Keras • Pytorch • Numpy • Pandas • Matplotlib • SciKit-Learn • OpenCV • PowerBI• LATEX• Jupyter Notebook Technologies and Domains: Machine Learning • Deep Learning • Natural Language Processing • Computer Vision • GenAI • LLMs • RAGs • Data Analysis • Data cleaning • Data visualization • Exploratory Data Analysis • Feature engineer- ing • Data Pre-processing • Probability • Hypothesis testing • Descriptive & Inferential statistics • Research • Documentation Education

Indian Institute of Science, Bangalore, India Jul 2019 PhD & MS in Civil Engineering

Indian Institute of Technology, Kharagpur, India Jun 2010 B.Tech (Hons.) & M.Tech in Mining Engineering CGPA: 8.14/10 Additional GPA: 9/10 Work Experience

Toshiba, Seattle, WA, USA Nov 2020 - Present

Senior Research Engineer(Applied AI)

• QA-bot domain customized

- Architected a Retrieval-Augmented Generation (RAG) pipeline to enhance the accuracy and relevance of AI- generated content by integrating both structured and unstructured data sources.

- Created a cutting-edge Question-Answering (QA) bot utilizing LLaMA-2, LangChain, and Streamlit to provide accurate responses by analyzing uploaded documents.

- Deployed end-to-end machine learning systems using technologies like Streamlit, Flask, and REST API for production environments, ensuring robust deployment of AI solutions.

• Arc quenches prevention in Electric Arc Furnaces

- Built predictive models to prevent arc-quenches in electric-arc furnaces of steel factory for TMEIC (Toshiba Mitsubishi- Electric Industrial Systems Corp)

- Employed machine learning classifiers such as Random Forest, XGBoost, Logistic Regression, SVM, Ensemble and K-means clustering methods to analyze data and build models.

- Performed Data pre-processing, Feature engineering, Hyper parameter tuning for models optimization.

- Analyzed data with visualization, discovered anomalous data and improved the overall data acquisition system

- Reduced operational shutdowns by 40%, Saving the costs associated with expensive restart operation of Electric Arc Furnaces that is 10% per each downtime.

• OCR for LATEX markup extraction

- Designed and implemented advanced OCR tool for extracting LATEX markup script from images with mathematical text such as equations and formulae.

- Created synthetic data images of mathematical text due to short fall of real dataset

- Employed image processing tools using OpenCV, image processing algorithms like character segmentation and Re- cursive Projection Profile Cutting (RPPC) algorithm for image pre-processing.

- Built image to sequence prediction models on CNN-RNN architecture employing loss function as Connectionist Tem- poral Classification (CTC) Loss.

- Converted mathematical text into readable LATEXscripts, streamlining data processing with 88% Accuracy and 1.4 Character Error Rate (CER).

• Forecast of Water Inflow in Dam

- Developed time series based forecasting models to predict water inflow in Dams of Japan.

- Developed statistical forecast models such as ARIMA, SARIMA and ML algorithms like LSTM, multi-LSTM leveraging weather forecasts and rainfall data for accurate inflow predictions.

- Improved the water resources management and planning aiding in effective dam operation and flood prevention. GoAir, Mumbai, India May 2020 - Oct 2020

Data Scientist

• Revenue Optimization

- Analyzed demand curves and implemented dynamic pricing algorithm to maximize the revenue.

• Consumer Behaviour

- Developed an NLP tool to calculate text similarity scores for a recommender system.

- Led a team of junior AI engineers and data analysts. CGI, Bengaluru, India Feb 2018 - Apr 2020

Lead Analyst

• Oil & Gas Exploration

- Built software add-ons for CoFlow Reservoir & Production software to estimate mechanical fiels such as subsidence and displacement.

- Implemented algorithms in C++ and Python with end-end focus from High-level design through Unit testing and Integration testing.

- Led a cross-functional team of 15, Played SME role

• Risk Analysis

- Designed POC for risk mitigation application that predicts time to failure of sand in oil wells

- Employed machine learning and geo-Statistics methods including Kriging, Regression and Bayesian statistics to correlate log data and core data and estimate the failure condition of rock IISc, Bengaluru, India Aug 2010 - Apr 2017

Senior Research Fellow

• Image Processing based Deformation Measurements

- Implemented image processing algorithms (PIV and PTV) in Matlab & C++ on images with high speed cameras.

- Measured velocity, displacement on granular flows. Argued internal circulation as reason for observed anomaly.

- Developed data driven constitutive models for sheared granular materials leveraging regression techniques.

- FFT algorithm on stress sensor data and discovered periodic stress fluctuations in slow granular flow. Academic Projects

• Mine Economics:

- Implemented Lane’s cut-off grade algorithm to schedule Mining operations - Optimized Net Present Value of mining plan by forecasting revenues and expenses. Implemnted algorithm in C and Python

• Rock Modelling:

- Developed empirical formulae for grouted rock bolt deflections in tunnels and validated with experimental data

• Modelling the self similarity of powders nano grinding:

- Implemented a software in C for nano-grinding model

- Validated a non-linear nano grinding model through experiments Publications

• Kumar, V.S., Murthy, T. and Nott, P.R., 2013, June. “Rheometry of dense granular materials: the crucial effects of gravity and confining walls”. In A. Yu, K. Dong, R. Yang and S. Luding eds.,AIP Conference Proceedings (Vol. 1542, No. 1, pp. 49-51). AIP.

• Kumar, V.S., and Murthy, T.G., “Effect of initial depositional fabric on the stress profile in a sheared granular column”, 8th International Conference for Conveying and Handling of Particulate Solids, CHOPS 2015 Dan Panorama Hotel, Tel Aviv Israel 3 May - 7 May 2015.

Honours and Awards

– Best Poster at COMPFLU(Complex Fluids) conference Dec 2014

– IISc Research fellowship 2010-2016

– GATE Fellowship 2009-2010

– All India Rank 5 in GATE exam 2009

– Merit-cum-Means scholarship 2005-2009

– GARP, IISc International travel grant 2015

– Top 1 % in IIT-JEE Mains 2005

– Top 0.5% in All Indian Engineering Entrance Exam 2005

Contact this candidate