Vasanth Dhanagopal Email & Phone Number

Chennai, Tamil Nadu, India

• Experimented with image semantic segmentation, utilizing models like ConvNext, DINO, MaskFormer,Mask2Former, SegFormer, and Vision Transformers (ViT).
• Constructed an instruction fine-tuned (decoder-based) text generation pipeline (GPT-3, BLOOM, facebook-OPT, Llama, Mistal) employing RAG with Langchain, and Qdrant Vector DB.
• Detected personally identifiable information using NER with encoder models such as BERT, DistillBERT, RoBERTa, DeBERTa.
• Applied quantization techniques (AWQ, GGML, bitsandbytes) & PEFT (LoRA), to achieve inference withlarge language models of upto 13B parameters.

Chennai, Tamil Nadu, India

• Performed tweet sentiment extraction & distinguished tweets related to real disasters, utilising MLclassification models for NLP.
• Implemented a multiclass image classification approach, utilizing pretrained models (ResNet, ImageNet,EfficientNet) and their ensembles on an extensive dataset of fundus photography images.
• Engaged in wheat head & pothole detection projects, employing image analysis with pre-trained models such as R-CNN, FPN, Retinanet, Yolo, and Efficient-Det.

Chennai, Tamil Nadu, India

• Bank Loan defaulter prediction was carried out with varied supervised algorithms (Logistic Regression, SVM, KNN, Decision Tree, Random Forest, ANN) for performance study.
• A cab booking interface embedded with weather, currency conversion using API calls was created in Python-Tkinter with login & registration pages along with database connectivity using MySQL server.
• Tasks in Python using functions, decision making statements, regular expressions along with file handling were developed and deployed using Flask API.
• Exploratory Data Analysis was performed on the pharmaceutical data for optimizing the drug stock management in order to reduce the bounce rate and time series models (ARIMA, LSTM) were applied.
• Hierarchical clustering and K-means techniques were framed for the airline's data; optimum number of clusters was obtained.

Chennai, Tamil Nadu, India

• Cross Industry Standard Process with monitoring and maintenance (CRISP-ML) is studied with various data processing techniques.
• Descriptive analysis is performed in finding the environmental impact for logistic industry.
• Customer engagement, sales trend, impact of discount of a Korean fashion in the Philippines on online platforms is reported using Tableau dashboards.
• Using exploratory data analysis, a linear regression model is constructed for predicting the particulate matter (PM2.5).
• Framed a weather analysis and forecasting report using Artificial Neural Network model for predicting snowfall.
• PCA is applied and principal components are selected with cumulative variance of 72.56% for the clinical data of patients pertaining to heart disease using AWS Sage-maker.

Toronto, Ontario, Canada

• Project: Reaction Wheel Failure of Kepler Spacecraft
• Identified key metrics and analysed the faults of reaction wheels of the spacecraft, using friction theory model and correlation techniques.
• Interpreted the telemetry data of the spacecraft using LSTM recurrent neural network.
• Created a prognostic model with heuristic optimisation tools (genetic algorithm and particle swarm optimization).Project: Remaining useful life prediction of Turbofan Engines
• Derived a prognosis algorithm with Savitzky-Golay filter and incorporated dimensional reduction for the same fleet of turbofan engines.
• Validated an enhanced statistical technique using neural network model for predicting the remaining useful life.Project: Boosting customer footfall for a retail store

Bengaluru, Karnataka, India

• Project: Performance enhancement of Military Aircraft
• Developed an in-house flow solver with a multi-blocking feature with 12% increased accuracy and 25% reduced computational time.
• Performed a time-based flutter analysis on an aircraft wing with 9% improved capability.
• Carried out a health monitoring test on a wing-body configuration for 0.7 & 0.8 Mach numbers.
• Generated unstructured meshes and studied the variations for multi-body configurations.

Bengaluru, Karnataka, India

• Project: Gas Turbine Engine Externals Optimisation
• Designed more than 75 tubes, end-fittings and structural components for various jet engines with 100% on-time delivery/detailed technical report.
• Proposed a solution for an oil leak problem of TRENT1000 engine in TEAM VIEWER.
• Rerouted a fire line of TRENT700, eliminating the hazard with 25% increased structural integrity.
• Developed schematics for external components and flow systems of TRENT1000 TEN engine.

Master'S Degree, Aerospace, Aeronautical And Astronautical/Space Engineering, 4.13/4.33

Activities and Societies: Invigilator, Teaching Assistant, Research AssistantResearch Fellow in Artificial Intelligence for Aerospace.

Bachelor'S Degree, Aerospace, Aeronautical And Astronautical/Space Engineering, 9.1/10

Activities and Societies: Departmental Event Organizer, Rotract Membero Designed and analyzed the effects of location and geometry of.