Vedika Srivastava

AI/ML Engineer

Work Experience

AI Research Scientist

Boston University - Boston Medical Center

Boston, MA, USA

Since September 2024

  • Leading AI-Driven Healthcare Research: Spearheading three concurrent AI research projects, developing deep learning solutions for medical diagnostics and predictive analytics.
  • EEG-Based Cerebral Edema Detection: Designed Transformer and LSTM-based deep learning models, achieving 92% test accuracy in early cerebral edema prediction using multi-channel EEG data.
  • CGM Prediction Model for Diabetes Management: Built a continuous glucose monitoring (CGM) prediction model, integrating meal intake data to provide real-time glycemic trend forecasts for diabetes patients.
  • GAN-Based Knowledge Graph Augmentation: Developing a GAN-driven knowledge graph augmentation system, improving data diversity and enhancing AI-assisted medical knowledge retrieval for clinical applications.
  • Mentorship & Compliance: Supervising 5 graduate students, providing hands-on mentorship in deep learning, clinical AI deployment, and data privacy while ensuring adherence to HIPAA regulations.

Agentic AI Engineer

Idea by Design Corp. (Stealth Startup)

Boston, USA

Since May 2024

  • Multi-Agent LLM Orchestration: Spearheading multi-agent LLM architecture development, enabling advanced AI-driven automation workflows with distributed model coordination.
  • Advanced AI System Architecture: Implementing scalable AI systems using OpenAI, Azure, and CosmosDB, optimizing real-time knowledge retrieval with vector databases (FAISS, ChromaDB).
  • Prompt Engineering & CI/CD Deployment: Enhancing LLM performance through advanced prompt engineering while setting up CI/CD pipelines for scalable AI deployment.

Lead AI Engineer

BeSpoke Care Inc. (Stealth Startup)

Boston, USA

June 2024 - February 2025

  • Conversational AI for Education: Engineered interactive storytelling AI with multimodal NLP and voice synthesis using GPT-4o Mini, AWS Polly, and Gemini, improving user engagement for children.
  • Multilingual & Adaptive AI Solutions: Integrated Google Polyglot, Dialogflow CX, and Ludwig, enabling real-time multilingual conversational AI and personalized, age-adaptive content recommendations.
  • Edge AI & IoT Implementation: Designed a multi-threaded AWS (EC2, Lambda) architecture to support Raspberry Pi/ESP32-based smart healthcare devices, improving dynamic content delivery for 100+ devices.

AI Engineer I

Time Machine Learning Inc. (Stealth Startup)

Boston, USA

March 2024 - June 2024

  • Advanced AI System Architecting: Engineered the VectorDatabase and RAG development utilizing all-mpnet-base-v2 for vector embeddings. My work in prompt engineering and MLOps led to a 25% improvement in system performance and scalability, setting new benchmarks in knowledge creation and operational efficiency.

Teaching Assistant

Boston University College of Arts and Science

Boston, USA

January 2023 - December 2023

  • Leadership in Educational Support: Excelled as a Teaching Assistant for "Computational Tools for Data Science," where I led over 30 labs and office hours weekly. My efforts were pivotal in developing the AI and Data Science skills of more than 200 students through hands-on training in Python, sklearn, TensorFlow, and PyTorch.
  • Academic Mentorship: Provided personalized mentorship and reviewed assignments on key machine learning algorithms such as K-means, KNN, SVM, and GMM. Offered specialized support on advanced topics including TF-IDF, Word2vec, and GANs, utilizing tools like NLTK, SpaCy, and Gensim to foster deep learning.
  • Recognition for Excellence: Received accolades from students and faculty alike for my dedication to enhancing the educational experience. My commitment was reflected in 95% positive evaluations, underscoring my role in significantly boosting students' competencies in data science.

Data Science Fellow

Institute of Global Sustainability

Boston, USA

May 2023 - July 2023

  • Worked in close coordination with NREL (National Renewable Energy Laboratory) advisors, benefiting from their guidance and expertise in the field.
  • Strategic Energy Planning: Spearheaded the development of a comprehensive energy planning tool, leveraging data from the RECS (Residential Energy Consumption Survey) and Justice40 datasets to inform sustainable energy policies and initiatives.
  • Collaborative Data Integration and Analysis: Analyzed and integrated data from RECS, Justice40, and LEAD datasets, unveiling insights that influenced the development of policies addressing energy inequalities and justice issues.
  • Innovative Data Visualization: Developed 20+ dynamic and interactive data visualizations using Plotly Dash, enhancing the tool's utility for strategic energy planning and policy making, and facilitating informed decision-making among stakeholders.
  • Research and Presentation: Successfully presented research findings to over 50 stakeholders, effectively communicating the critical implications of sustainable energy planning and policy based on comprehensive data analysis.

SDE Research Intern

IBM Integrated Software Labs

Pune, IN

January 2022 - July 2022

  • Framework Optimization and Migration: Led the successful migration and optimization of the Quarkus framework for an IBM Java-based microservice, achieving a notable 50% reduction in resource usage and a 13% increase in overall efficiency.
  • Product Enhancement through Cross-Functional Collaboration: Coordinated with a cross-functional team in Singapore to upgrade IBM Security Verify, achieving a 50% reduction in resource usage and a 13% increase in operational efficiency, directly contributing to enhanced product marketability and customer satisfaction.
  • Led efforts to optimize system performance and scalability. These improvements contributed to higher customer satisfaction and cost savings for clients.

NLP Research Intern

Tech Mahindra Maker's Lab

Pune, IN

October 2020 - January 2021

  • Pioneering Voice Bot Development: Developed a novel voice bot in Sanskrit, integrating data analysis and web scraping across 20+ data sources with MySQL database operations. This groundbreaking project demonstrated the potential for leveraging technology to preserve and promote linguistic diversity.
  • Algorithmic Excellence and Backend Optimization: Conducted comprehensive research on over 11 algorithms, optimizing backend functionalities for enhanced performance. The project showcased substantial improvements in model accuracy and scalability, contributing to over 1500+ languages.
  • Contributed to over 1500+ languages, showcasing substantial improvements in model accuracy and scalability.