Hi! My name is Devi Prasad, a Machine Learning Software Engineer, Full Stack Developer.
My specialization
What i do
Deep Learning
CNN (DenseNet, ResNet, Inception), FCN, RNN (GRU, LSTM, bi-LSTM,
stacked-LSTM, VAEs, Object Detection, Transformers, Transfer
Learning, customizing, debugging and tuning Neural Networks.
Machine Learning
Linear Regression, logistic Regression, KNN, Naive Bayes, SVM,
RL, Bagging & Boosting (ensemble), PCA, SVD,regularization,
accuracy measures, statistical analysis and modeling, and
knowledge of end-to-end machine learning pipeline (feature
engineering, data cleaning, data visualization, model deployment
and serving)
Software Development
Data structures and Algorithm, C++, Python, Java, GO Rust,
Swift, Pytorch, Tensorflow, FastAI, Scikit-Learn, OpenCV, CUDA,
HTML, CSS, JS.
Scrum, Agile process, UML, Build Pipeline, Junit-test, unit
test, Java Docs, Doxygen, Build tools (Cmake, gcc, Make, gmake).
Full stack development: RestAPI, GRPC, Flask, NodeJS.
Mobile Development: Android (Java, Kotline), Flutter
(Dart), Xcode (Swift), Figma, Adobe XD, Human Computer
Intreaction Certified.
Cloud Computing/Deployment
AWS Certified, Kubernetes cretified, Google Cloud platform,
Azure.
Experience & skills
EXPERIENCES
Alectio
- Created Alectio’s infrastructure on Kubernetes with OpenShift, Kops, Terraform, Nginx-Ingress, Kubeflow, AWS Cloudfront, AWS VPC, Cloud Trail, and Airflow.
- Re-designed and implemented Alectio’s backend in Golang and migrated Dynamodb database to Cassandra. Created Dev environment setup using VSCode dev container, used goSwagger to manage and automate lots of code management.
- Redesigned Alectio’s Client python SDK to work with GraphQL and use GRPC for an end to end communication with Alectio’s server
- Build hosted Auto Active learning stack with ephemeral compute instances using Kubernetes. Handled provisioning, deletion, data retrieval, deployment monitoring for the training instance by writing infrastructure as code backend to interact with various infrastructure resources.
- Built Auto labeling stack and deployed using KubeFlow (Tensorflow extended)
- Built the ETL stack for run analysis on high volume streaming data using Spark and Hadoop
- Built stack for internal instrumentation using Grafana, Python Flask, and Postgres SQL
- Built CI/CD pipeline using Jenkins and Gitlab to automate testing of APIs, log changes to Confluence and Receive errors messages
Third Insight
From developing deep learning based various components of
visual scene understanding deep learning pipeline to
building C++ and CUDA based systems, I had a phenomenal
experience of wearing a lots of hats and taking ownership.
ThermoFisher Scientific
Implemented and trained CNN based models for next-generation
QPCR and PCR analysis software using Python and Tensorflow.
Improved existing system accuracy from 78% to 98%. Solved
multi dye cross-talk problems. Developed various
methodologies to evaluate the performance of the model
according to the domain. Built front-end system to use the
Genotyper
BNSF
During the 3 months internship, got the pleasure of building
a computer vision application to detect defects in the
railway tracks and developing various data analysis
algorithms.
Urstuffmyhouse
In order to help the community with its need for reliable
storage, we built a MVP and launched it to aid people in
finding a place to store their belongings safely and
conveniently
Thank you for visiting my site. Let's talk more.
Contact
10k
1
15
.svg){kind=small}