M
A
N
O
H
A
R
V
E
L
L
A
L
A
.M
S
C
S
@
O
D
U
R
e
s
e
a
r
c
h
A
s
s
i
s
t
a
n
t
As a driven Master of Computer Science student with experience in software development, I embarked on my coding journey when my cousin Mahesh Avadhanam, introduced me to his work and asked me to check out AlgoExpert.io.
This platform was instrumental in honing my coding skills, laying a solid foundation in Data Structures and Algorithms , Which enabled me to support the development of web applications using tools like React, Spring-boot, Flask, Django in large organizations.
Tech Stack
Aug. 2023 - Present
- Revamped Map Communications Inc. login portal using AWS Lambda , enabling over 100K+ clients to securely access the portal. Utilized OAuth 2.0 and Twilio SMS API for sending OTPs.
- Designed CI/CD pipelines using Jenkins to automate building, testing, and deployment of Django, React, MongoDB based virtual receptionist system for Map Communications Inc.
- Engineered the VSortsTM cloud-based SaaSweb app with cross-functional team (Product, Sales, Engineer, Support, Designers) of 5. Constructed reusable components using Material UI, Tailwind CSS.
Feb. 2023 - Jun. 2023
- Improved internal access control turnaround time and security by developing a module using Java Spring-Boot, integrated with OAuth 2.0 for authentication and authorization, automating access control.
- Increased productivity by reducing the time and effort required to discover and access data, by deploying a data discovery tool Amundsen to a Kubernetes (Amazon EKS) cluster, leading to more informed decision-making.
Aug. 2021 - Oct. 2021
- Designed, developed, and maintained the Flask-based Electronic Beam Software web application for I.S.R.O
- Introduced a Logistic Regression ML model and minimized the Cross-Entropy Loss to 2% using Gradient Descent method to intelligently update the coefficients, Predicting the intensity currents in the welding process.
- Applied L1 regularization to penalize large coefficients and prevent overfitting of the Model which helped in detecting anomalies in the welding process, saving 120+ Engineer hours.
- Innovated CS120.AI an Angular, Django based chatbot designed for courses at Old Dominion University.
- Harnessed the power of the Transformers from Hugging Face to fine-tune the BERT model and developed APIs to seamlessly integrate it into CS120.AI serving as a Large Language Model (LLM).
Developed Convolutional Neural Network using TensorFlow 2 Keras API for classifying Areca Nut Images as Healthy or Diseased. Employed a CNN architecture with eight layers, including Convolution layers with ReLU activation, max pooling layers, and a final dense layer for prediction.
- Conducted in-depth analysis of Diabetes Survey dataset, employing PCA and feature engineering methods to predict diabetes based on factors such as Blood Pressure, BMI, and Pregnancy
- Implemented Supervised Learning models, including SVM, Logistic Regression, and Decision Tree . Identified and addressed outliers within clusters, employing Soft-Margin SVM with Slack to overcome constraints.
Developed Graph Visualizer with Node.JS to comprehensively visualize Graph Algorithms, including SWARM, Dijkstra’s, BFS, and DFS. The Website is deployed live at: Graph Visualizer.
