Sri Vishnu R

Machine Learning Engineer

[email protected]||+1 (940) 299-8408

PROFESSIONAL SUMMARY

 Accomplished Machine Learning Engineer and Data Scientist with over 12 years of diverse IT
experience, including more than 8 years in designing, developing, deploying, and monitoring advanced
ML/AI solutions.
 Adept in architecting end-to-end machine learning systems from data collection and preprocessing to
training, deployment, and production monitoring using best practices in MLOps and DevOps.
 Proficient in building and maintaining automated ML pipelines using Airflow, MLflow, SageMaker
Pipelines, TFX, and Kubeflow to enable reproducibility, versioning, and CI/CD for ML models.
 Deep experience implementing model monitoring and retraining workflows using tools like
Prometheus, Grafana, EvidentlyAI, and AWS CloudWatch, ensuring model performance over time.
 Hands-on expertise with cloud platforms: AWS (SageMaker, Lambda, ECS, S3, Redshift), Azure (ML
Studio, Synapse), and GCP (Vertex AI, BigQuery) – including hybrid and multi-cloud architectures.
 Strong background in supervised, unsupervised, and semi-supervised learning, with practical
application of algorithms like Random Forests, XGBoost, LightGBM, Gradient Boosting, SVM, and
KNN.
 Proficient in deep learning techniques, using PyTorch, TensorFlow, and Keras to build models like
CNNs for image classification, LSTM/GRU/RNN for sequence modeling, and Transformers (BERT,
T5) for NLP tasks.
 Developed and deployed multiple Natural Language Processing (NLP) solutions, including sentiment
analysis, topic modeling, text summarization, chatbot systems, and information extraction pipelines.
 Designed and productionized computer vision solutions for image segmentation, defect detection,
object tracking, OCR, and image-based recommendation using OpenCV, YOLOv5, Bytetrack, and
Detectron2.
 Built and scaled recommendation systems (collaborative filtering, content-based, and hybrid), fraud
detection systems, and churn prediction models for Fortune 500 clients across sectors.
 Experience in stream processing frameworks such as Apache Kafka, Apache Flink, and Spark
Streaming to support real-time ML inference and low-latency decision systems.
 Created data lakes and lakehouses on AWS S3/Glue and Azure Data Lake with Delta Lake to store
petabyte-scale raw, processed, and feature-engineered data efficiently.
 Led the design of feature stores to standardize feature engineering and sharing across teams using
Feast, Tecton, and custom-built in-house platforms.
 Skilled in data wrangling, transformation, and ETL/ELT workflows using Spark (PySpark/Scala),
Hive, Airflow, and Databricks, working with structured and semi-structured data (JSON, Parquet,
Avro).
 Conducted A/B testing and statistical analysis of ML models and business experiments using Bayesian
methods, t-tests, ANOVA, and uplift modeling.
 Created comprehensive dashboards and visualizations using Power BI, Tableau, and Plotly Dash to
track model metrics, business KPIs, and data quality across various pipelines.
 Strong expertise in model interpretability and fairness using SHAP, LIME, Captum, and Fairlearn to
support ethical AI practices and transparency for regulated industries.
 Developed containerized ML services with Docker and Kubernetes, enabling scalable and resilient
deployment of inference endpoints and batch scoring jobs.
 Experience building APIs with FastAPI, Flask, and gRPC to serve machine learning models with
latency and throughput constraints in production environments.
 Practiced in MLOps lifecycle, maintaining complete traceability of data, code, parameters, and models
using DVC, Git, MLflow, Weights & Biases, and other versioning tools.
 Familiar with Data Governance, PII handling, GDPR, HIPAA, and responsible AI practices, especially
in healthcare and finance sectors.
 Excellent communication, documentation, and client engagement skills; have worked closely with
stakeholders to translate vague business goals into measurable ML deliverables.
 Actively mentor junior data scientists and engineers, contribute to code reviews, and participate in
internal AI/ML communities of practice to promote knowledge sharing.

TECHINALS SKILLS

Languages Python, R, Scala, Java, SQL, PL/SQL, Shell Scripting, C, C++

Framework Flask, FastAPI, Django (basic), Apache Airflow, Luigi, TFX, Spark MLlib.

ML Libraries Scikit-learn, Pandas, NumPy, Matplotlib, Seaborn, XGBoost, LightGBM, Statsmodels,

SciPy, CatBoost
Cloud Platforms AWS (S3, EC2, SageMaker, IAM, Lambda, CloudWatch, Redshift, Glue, Athena,
ECR, ECS, CodePipeline), Azure (ML Studio, Synapse), GCP (Vertex AI)
Deep Learning / AI TensorFlow, PyTorch, Keras, OpenCV, HuggingFace Transformers, Bytetrack,
Detectron2, ONNX
NLP Tools SpaCy, NLTK, Gensim, TextBlob, BERT, GPT, T5, LLaMA, LangChain

Computer Vision OpenCV, YOLOv5/v8, Bytetrack, Detectron2, MMDetection, ImageAI

MLOps Tools MLflow, DVC, Weights & Biases, TensorBoard, Neptune.ai, Feast, Tecton, Kubeflow
(basic), Metaflow, Dagshub.
Version Control Git, GitHub, GitLab, Bitbucket

Containerization Docker, Docker Compose, Kubernetes (basic), Amazon ECS

CI/CD Jenkins, GitHub Actions, GitLab CI/CD, Argo Workflows, CodePipeline, Airflow
CI/CD
Monitoring & Logging Prometheus, Grafana, ELK Stack (Elasticsearch, Logstash, Kibana), AWS
CloudWatch
Databases PostgreSQL, MySQL, Oracle, SQL Server, MongoDB, DynamoDB, Snowflake,
Redshift
IDE & Notebooks JupyterLab, Jupyter Notebook, RStudio, PyCharm, VS Code, Notepad++.

Visualization Tools Tableau, Power BI, Plotly, Dash, Matplotlib, Seaborn

Statistical Analysis R, Excel, Statsmodels, Bayesian Inference, A/B Testing, ANOVA, Uplift Modeling
 PROFESSIONAL EXPERIENCE

Client: Engie Impact Jan 2023 to Till Date

Role: MLOps Engineer

Responsibilities
 Designed and deployed a machine learning pipeline to automate price prediction for retail products
across multiple geographic markets.
 The solution was built for a large retail client looking to optimize pricing based on demand forecasts
and historical sales data.
 Developed Python-based REST APIs using Flask for model inference endpoints.
 Orchestrated ETL and model training pipelines using Apache Airflow integrated with MLflow.
 Implemented model versioning and tracking using MLflow on AWS, with models deployed via S3
and EC2-based Flask servers.
 Used Data Version Control (DVC) and Dagshub for managing datasets and pipeline reproducibility.
 Containerized applications using Docker to ensure environment consistency from dev to prod.
 Created automated monitoring dashboards using Grafana to track pipeline health and performance.
 Integrated CI/CD with GitHub Actions for automated testing and deployment of new model versions.
 Implemented end-to-end CI/CD pipelines using Jenkins, Git, and Ansible, streamlining the software
delivery process and reducing time to market.
 Designed and deployed scalable and secure cloud infrastructure on AWS, leveraging services such as
EC2, S3, and RDS for reliable and high-performing systems.
 Containerized applications using Docker, facilitating consistent deployments and efficient resource
utilization.
 Orchestrated containerized applications using Kubernetes, ensuring high availability, scalability, and
automated deployments.
 Implemented configuration management using Ansible, ensuring consistent and reproducible
infrastructure configurations across environments.

Client: E-commerce Company Jun 2021 to dec 2022

Role: Machine Learning Engineer
Responsibilities
 Designed, implemented, and deployed a scalable multi-model ML forecasting engine for real-time
inventory optimization across 30+ regional fulfillment centers, reducing out-of-stock incidents by 37% and
cutting excess inventory costs by $5.6M annually.
 Developed end-to-end ML pipelines with Apache Airflow, Docker, MLflow, and integrated with AWS
SageMaker Pipelines, automating retraining, model versioning, and zero-downtime deployments triggered
by event-based AWS Lambda functions.
 Engineered a hybrid recommendation engine combining collaborative filtering, content-based filtering, and
Neural Matrix Factorization, resulting in a 22% boost in user session duration and a 14% lift in cross-sell
revenue.
 Led an NLP initiative to mine customer feedback using Transformer-based models (BERT, RoBERTa) for
aspect-based sentiment analysis, helping drive key product roadmap decisions and improve CSAT scores
by 18%.
 Collaborated with Data Engineering to architect a cloud-native data lakehouse leveraging AWS Glue,
Athena, S3, Redshift Spectrum, and Delta Lake format for high-performance ad-hoc analytics and ML
training data pipelines.
 Built robust fraud detection and credit scoring systems using XGBoost, LightGBM, and stacked ensemble
architectures, raising production AUC from 0.72 to 0.88, decreasing false positives by 31%, and catching
$4M+ in fraud annually.
 Developed a feature store with point-in-time correctness using Feast and AWS DynamoDB, serving both
batch and online features to downstream models in under 50ms latency.
 Created real-time data ingestion pipelines using Kafka, Spark Streaming, and AWS Kinesis for monitoring
supply chain anomalies, triggering proactive alerts via SNS and Lambda.
 Used SHAP, LIME, and Fairlearn to audit model fairness and interpretability in credit models to comply
with GDPR and FCRA regulations.
 Deployed multi-armed bandit experimentation framework for A/B testing different recommendation
strategies, resulting in a 6.8% uplift in CTR with significantly lower sample sizes.
 Conducted regular model drift analysis and data integrity checks using Evidently AI, improving long-term
model performance consistency by 19% over 12 months.
 Presented quarterly insights to senior leadership with interactive dashboards (Tableau, Power BI) and
custom Dash apps integrated with live data from production APIs.
 Built and automated customer churn models using survival analysis and Gradient Boosting, enabling
personalized retention offers and improving retention by 9% quarter over quarter.
 Worked on anomaly detection in high-velocity clickstream data using Isolation Forest and AutoEncoders,
identifying bot traffic and behavioral anomalies within seconds.
 Played a key role in designing internal ML platform tooling to support CI/CD of models, automated
retraining workflows, metadata tracking, and governance.

Client:- Global Healthcare Services Jun 2018-May 2021

Role:- MLOps Architect
Responsibilities
 Designed and deployed patient risk stratification models using structured claims, unstructured clinical
notes, and EHR data. Applied Random Forest, Logistic Regression, and XGBoost for classification and
ranked risk scores; resulted in a 26% improvement in early intervention effectiveness.
 Engineered deep learning architectures using LSTM, CNN, and autoencoders for longitudinal patient
health data (e.g., vitals, labs, medication adherence) to predict hospital readmissions, achieving 81%
precision and 76% recall.
 Developed a chronic disease progression model leveraging temporal convolutional networks (TCN) to
forecast comorbidity emergence (e.g., diabetes → renal complications), enabling tailored care plans and
improved treatment adherence.
 Spearheaded the implementation of MLOps best practices with Kubeflow, TensorFlow Extended
(TFX), Docker, and Cloud Build pipelines on GCP AI Platform, cutting model deployment time from
weeks to under 2 days.
 Integrated model explainability tools (SHAP, What-If Tool) to enhance clinical trust and auditability,
and mapped outputs to HL7/FHIR-compliant APIs for EHR system integration.
 Developed and productionized fraudulent billing detection systems in real-time using Isolation Forests,
Variational AutoEncoders, and Dynamic Time Warping (DTW), reducing false positives by 45% and
recovering $2.3M/year in detected fraud.
 Deployed predictive care pathway models that dynamically recommended next best clinical actions,
leading to a 1.3-day reduction in average patient stay and an 11% improvement in bed utilization rates.
 Created data pipelines using GCP Dataflow, BigQuery, Cloud Functions, and Cloud Scheduler to
ingest, clean, and process petabyte-scale structured and semi-structured clinical datasets with low-latency
SLAs.
 Designed an internal Model Governance and Metadata Tracking System for model lineage, versioning,
and rollback using ML Metadata (MLMD) and Kubeflow Pipelines, improving auditability and
reproducibility.

Client:- Retail Client apr 2015-May 2018

Role:- Big Data Engineer
Responsibilities
 Architected and maintained highly scalable ETL pipelines processing over 10 million+ daily customer
transactions and product events, leveraging Apache Spark and MapReduce frameworks to ingest,
transform, and store data into HDFS, Cassandra, and Hive clusters with near real-time SLA compliance.
 Designed and optimized Spark SQL jobs and Hive queries to support large-scale analytics for
merchandising and inventory teams, enabling dynamic pricing and stock replenishment strategies that
increased inventory turnover by 15%.
 Built end-to-end real-time streaming pipelines using Apache Kafka as the message bus and Spark
Streaming for processing, enabling real-time customer segmentation and targeted marketing campaigns
during high-impact events like flash sales and holiday promotions, improving conversion rates by 12%.
 Developed NLP pipelines using word2vec embeddings and supervised text classification models (SVM,
Random Forest) on millions of product reviews and customer feedback, improving automated catalog
tagging, faceted search relevance, and personalized recommendations.
 Partnered with cross-functional teams to design and implement a feature store architecture, enabling
consistent feature reuse and governance across multiple machine learning projects, reducing feature
engineering time by 40% and minimizing data inconsistencies.
 Conducted advanced data profiling and anomaly detection on streaming and batch data, using Spark
MLlib to flag data quality issues early, improving pipeline reliability and reducing downstream error rates
by 30%.
 Developed interactive and executive-level predictive analytics dashboards using Tableau, integrating
retail KPIs like sales velocity, demand forecasting, and price elasticity to empower data-driven decisions at
the merchandising and operations levels.
 Collaborated with data scientists, engineers, and business analysts to translate complex retail business
questions into scalable big data solutions, facilitating predictive inventory replenishment and dynamic
pricing optimization.
 Automated cluster resource monitoring and tuning for Spark and Cassandra, achieving a 20% reduction in
cloud compute costs while maintaining throughput and reliability.

Client:- Telecom Client jul 2013 – mar 2015

Role:- Data Engineer
Responsibilities
 Managed and maintained enterprise-grade Hadoop clusters consisting of over 300 nodes across on-
premise and hybrid cloud environments, ensuring 99.9% uptime and high availability for mission-critical
telecom data pipelines.
 Designed and implemented Sqoop workflows to ingest structured relational data from Oracle and
Teradata into HDFS, enabling seamless batch analytics and reducing manual data preparation efforts by
70%.
 Developed and orchestrated end-to-end MapReduce, Hive, and Pig jobs for Call Detail Record (CDR)
analytics, churn prediction, and revenue assurance initiatives, resulting in improved customer retention
strategies and fraud detection accuracy.
 Built and maintained Oozie workflows to automate daily ETL processes and orchestrate multi-stage data
pipelines, improving pipeline stability and reducing manual intervention.
 Tuned and optimized Spark and MapReduce jobs by adjusting YARN container allocations, parallelism,
and memory settings, achieving 45% reduction in batch processing times and significant cost savings in
compute usage.
 Collaborated with domain experts and business analysts to create data marts for usage pattern recognition,
dropped call analysis, and proactive service quality metrics.
 Implemented HDFS data lifecycle policies and archival strategies to manage growing storage footprint
while ensuring compliance with retention and security standards.