Hi, I am Souradip !!

I am working as a Software Engineer II at SLB in the Industrial IoT Applications team. I completed my master's in Electrical and Computer Engineer from Purdue University (West Lafayette) in 2024 with a focus on Artificial Intelligence. Earlier, I did my bachelor's from Indian Institute of Technology Guwahati in 2019 majoring in Electronics and Communication Engineering. My passion is Computer Science. I am currently interested in building full-stack ML applications and software solutions involving Deep Learning, Generative AI, Agentic AI, LLMs, Cloud Computing and Optimization. Previously, I have worked as a software developer at Adobe Inc. India in the Adobe Commerce Platform team.

Education

Purdue University logo

Purdue University - West Lafayette

Degree: Master of Science (August 2022 - May 2024)

Major: Electrical & Computer Engineering

  • Computational Models and Methods
  • Introduction to Artifical Intelligence
  • Advanced Software Engineering
  • Random Variables and Signals
  • Deep Learning & Applications
  • Reinforcement Learning
Indian Institute of Technology Guwahati logo

Indian Institute of Technology Guwahati

Degree: Bachelor of Technology (July 2015 - June 2019)

Major: Electronics & Communication Engineering

  • Data Structures and Algorithms
  • Probability and Random Processes
  • Pattern Recognition and Machine Learning
  • Introduction to Parallel Computing
  • Optimization Methods
  • Speech Technology

Experience

Software Engineer - SLB (Schlumberger)

(June 2024 - Present)

I am currently working as a Software Engineer II at SLB in the Industrial IoT Applications team in Houston, Texas where I am involved in the development and maintenance of the full-stack application called ValveCommander that provides real-time and remote automation to control valves in the oilfield.
SLB logo

Software Engineering Intern - SLB (Schlumberger)

(May 2023 - August 2023)

I worked as a Software Engineering Intern at SLB in the Pressure, Pumping, and Chemistry Unit in Sugar Land, Texas where I developed a light-weight data simulator for automation testing of a well-site orchestration platform using .NET Core. The orchestration platform provides real-time and look-ahead visibility to help coordinate different stages of wellsite activities thus reducing non-pumping time. I created a CLI and a web UI in Angular to run simulations based on data generated by user-defined simulation parameters to reduce the cost of using hardware simulators for testing.

Software Development Engineer - Adobe Inc. India

(July 2019 - July 2022)

I worked as a Software Development Engineer I and II at Adobe Inc. India (Bangalore). Here, I was involved in the design, development and maintenance of java based micro-services under a cloud-based internal e-commerce platform used for creation and management of offers or products released/to be released to market. My primary area of expertise was in the prepurchase section of the transactional journey which includes merchandising, pricing and other related components of the product offering. I have contributed in multiple business initiatives and worked on areas of data pipeline automation in product cataloguing, business rule management system, enhancing catalogue search using natural language processing etc.
Adobe logo

Summer Research Intern - Samsung R&D Institute India

(May 2018 - July 2018)

I attended Samsung R&D Institute India (Bangalore) in 2018 as a summer research intern where I worked at the Advanced Technology Lab. Here, I worked on a deep learning model for predicting the stress on a glass surface of a mobile device based on the different 3D parameters of the glass (thickness, radius of curvature, etc). The model was based on the state-of-the-art PointNet architecture by Charles R. Qi et. al. The model was trained using transfer learning and modified to predict stress data in point cloud form obtained from Four-Point Bending simulations of glass in CAD.
Samsung logo

Summer Fellowship - FOSSEE, IIT Bombay

(May 2017 - July 2017)

I worked as an Software intern in the summer of 2017 at the open source organization FOSSEE (Free/Libre and Open Source Software for Education) based at Indian Institute of Technology Bombay. The organization is part of the National Mission on Education through Information and Communication Technology (NMEICT), which promotes the use of open source software tools to improve the quality of education and research. I worked on building software packages for a simulation tool called OpenModelica which helps in interfacing electrical circuits with Arduino UNO boards. I also provided GUI based modelling and simulation support for different types of AVR architectures.
FOSSEE logo

Research

Time-Constrained Recommendations: Reinforcement Learning Strategies for E-Commerce

Author(s): Sayak Chakrabarty, Souradip Pal
This work studies time-constrained slate recommendation via reinforcement learning and reproduces experiments for e-commerce using SARSA and Q-learning. It provides dataset setup, training, evaluation, and simulation strategies for Alibaba re-ranking benchmarks.

Budget-Constrained RLRS title graphicSimulation results

MM-PoE: Multiple Choice Reasoning via. Process of Elimination using Multi-Modal Models

Author(s): Sayak Chakrabarty, Souradip Pal
This study introduces a two-step scoring system to predict answers by eliminating incorrect options in multiple-choice visual reasoning tasks. It also investigates the performance of open-source multi-modal models like GIT, BLIP etc. on benchmark datasets like VQA,ScienceQA and AI2D in both single-shot and few-shot settings.

MM-PoE visual-reasoning workflowMM-PoE command-line interface

ReadmeReady: Free and Customizable Code Documentation with LLMs - A Fine-Tuning Approach

Author(s): Sayak Chakrabarty, Souradip Pal
This research presents a large language model (LLM)-based application that developers can use as a support tool to generate basic documentation for any publicly available source code. The tool is a RAG application built on LangChain using HNSW similarity search algorithm and also allows custom code data generation and QLoRA techniques to support easy fine-tuning of open-source LLMs like LLama2 and Gemma.

ReadMeReady retrieval-augmented generation workflow

Triple Pendulum Based Nonlinear Chaos Generator and its Applications in Cryptography

Author(s): Bikram Paul, Souradip Pal, Abhishek Agrawal, Dr. Gaurav Trivedi
This research explores post-quantum cryptography methods and proposes a novel approach to design an encryption scheme based on the chaotic dynamic physical system, which is derived from a mechanical model of a triple-pendulum system depicting nonlinear dynamics. The proposed cryptography scheme exhibits resistance against various attacks and is validated using benchmark tests, such as Lyapunov exponents test, bifurcation diagrams, sensitivity to parametric and to initial values, ergodicity, collision test, NIST, diehard randomness test etc. The proposed algorithm is implemented on an FPGA using System-Verilog.

Triple-pendulum chaos cryptography visualization

Projects

Testing patterns analysis visualization

Analysis of Testing Patterns For PTM-Enabled Open Source Software Projects

This project shows a study of the common testing patterns (unit, integration, e2e) involved in open source software projects that uses Pre-Trained machine learning models.

Decision Transformer model visualization

Reinforcement Learning via. Sequence Modeling using Decision Transformer

This project involved studying the effect of context length on the variances of the expected returns of the Decision Transformer model in OpenAI Gym environments and Atari games in both online and offline settings.

KL loss object-detection result

Box Regression using KL Loss

This project includes several experiments to reproduce the results of the paper "Bounding Box Regression with Uncertainty for Accurate Object Detection" by He et al. The paper introduces a new regression loss function called KL-Loss for accurately predicting the bounding box locations for object detection using localization variances. The methods in the paper were reimplemented in PyTorch and tested on PASCAL-VOC dataset.

Speaker-recognition result visualization

Automatic Speech Segmentation

The aim of this project was to segment speech sequences based on speaker transitions, where the number of speakers is not known beforehand. Additionally, it can identify the number of speakers along with the zones where single or multiple speakers are active. Both supervised and unsupervised diarization approaches on LPC(Linear Predictive Coding) features were performed and tested on synthesized audio clips.

OpenModelica hardware-in-the-loop simulation

OpenModelica HIL Simulation

The problem involves building a hardware-in-the-loop simulation with a DC motor as a software plant, and a proportional control algorithm running on the Arduino. The process is expected to run in real time which means that a delay should not be used. An OpenModelica package was built which is able to conduct such Arduino involving HIL simulation using interprocessing communication.

Skills

Languages

Web Technologies

Libraries & Frameworks

Hugging Face Weights & Biases LangChain

Software Tools

Elasticsearch

Contact

If you have any questions, or if you just want to say hi, please feel free to send a message or email me.
To connect with me, please make sure to follow me in any of the social media links below.