Experience

Scale AI

June 2025 - Dec 2025

GenAI Advisor Intern

  • Prompt Engineering: Enhanced model output quality by identifying failure modes in large language models and engineering optimized prompts to address weaknesses in complex reasoning, edge cases, and factual grounding.
  • Key Learnings: Training frontier LLM models to perform complex reasoning tasks using PEFT and LORA.

Jackal Tech Ltd

May 2025 - Aug 2025

Software Engineer Intern

  • Team Leader: Led a team of 4 developers to design and build a full-stack web application using React.js and Node.js, following Agile development practices.
  • Team Work: Collaborated with senior engineers and project managers to define system architecture, design RESTful APIs, and integrate backend services, contributing to a platform used by 100+ internal users.

Credit Suisse (UBS group)

June 2022 – July 2024

Technical Analyst

  • Lead developer and SME of the web service responsible for pretrade analytics used by Equity traders and portfolio managers and internal clients to analyze the impact cost, historical performance, positional analysis of an equity portfolio.
  • Refined monolithic architecture to loosely couple the services leading to a 15 % increase in frequency of deployment and reduction in bugs introduced.
  • Optimized the deployment of backend services across the CS infrastructure, eliminating dependency on obsolete hardware and reducing TCO by USD 125K.
  • Worked on ETL, Scorecard and Stitching processes to transform and load market trading data for Transaction Cost Analysis and Algorithmic Trading.

Cognizant

February 2022 - May 2022

Software Developer Intern

  • Built a solid foundation in Java by solving algorithmic problems, implementing core OOP concepts, and completing extensive coding exercises and assessments.
  • Strengthened problem-solving and analytical skills through recurring technical evaluations and one-on-one code reviews focused on Java-based solutions.

Publications

Artificial Intelligence and Motives of Murder-Based Prediction in India

SCOPUS-indexed journal - CIMS (Computer Integrated Manufacturing Systems), November 2023

  • This paper presents a viable option of AI decision-making by raising questions about an individual's accountability in relation to criminal conduct.
  • Automated Decision Making refers to the use of artificial intelligence and related technologies to assist or replace human decision-making in the criminal justice system.
  • AI is already in practice to assist with tasks such as predicting recidivism, identifying suspects, or determining sentences.
  • This paper uses an AI model to predict whether a particular state/UT is at risk of increase in the crime rate or not.

Projects

Smart Waste Management System Using Object Detection and Classification


  • Designed and implemented a machine learning system to detect and classify waste into categories: recyclable, organic, and non-recyclable.
  • Achieved this by using object detection algorithms and fine-tuned models for 92% classification accuracy.

Multi-Modal AI document Generator


  • Developed an application to generate a Grant proposal, Blog, Powerpoint presentation and Social media posts.
  • Reduced the time of report generation from 30 days to one or two days.

MedTeller: Vision-Language Transformer for Automated Radiology Report Generation


  • Developed MedTeller, a Vision-Language Transformer system for automated radiology report generation from chest X-ray images, combining ViT-Base and GPT-2 with cross-attention mechanisms to bridge visual and textual modalities.
  • Evaluated model performance using BLEU and ROUGE-L metrics.

MindEase - Emotion Detection Chatbot


  • Developed a chatbot using FastAPI and DistilBERT (Hugging Face Transformers) to detect user emotions from text and respond empathetically.
  • Containerized the entire backend with Docker and deployed it to an AWS EC2 instance with public access.
  • Implemented RESTful APIs to handle user queries and return emotion-tagged responses.
  • Tracked training and performance using ClearML, managing multiple runs and hyperparameter tuning.

Bitcoin Price Prediction


  • Implemented and compared multiple models, including Linear Regression, Random Forest, and XGBoost.
  • Achieved the lowest Mean Squared Error and highest R2 score of 98.5% with Linear Regression, demonstrating the model's efficiency for price prediction..

Employee Information Application


  • Developed a full-stack application to manage employee information
  • It enables HR departments to efficiently handle employee data and improve workflow processes.

Skills

Programming Languages

Python
90%
Java
85%
JavaScript
80%
HTML5
85%

Machine Learning & AI

TensorFlow
85%
PyTorch
80%
Scikit-Learn
90%
Computer Vision
75%

Databases & Cloud

MySQL
85%
AWS
75%
Docker
80%
Oracle
70%

Tools & Frameworks

Git
90%
Spring Boot
80%
Angular
75%
Kubernetes
65%

Contact

My Address

College Park, MD 20740

United States

Social Profiles

Email

rishabhranka9324@gmail.com

Available For

Full-time Opportunities

Internships