Data Engineering Intern candidate with AI/ML PhD

Kartika Makkar

*************@*****.*** +1-650-***-**** Sunnyvale, CA www.linkedin.com/in/kartikamakkar https://github.com/kartikasoni SUMMARY

AI Engineer and Computer Science PhD with a focus on deploying real-world AI systems. Developed a RAG-based AI tutor bot that improved answer relevance by 30% and reduced query latency by 40%. Engineered a multimodal healthcare consultant using LangChain and FastAPI for real-time clinical query handling. Contributed to a BiLSTM encoder-decoder model, achieving a 12.6% ROUGE score improvement for sarcasm interpretation, which was published in Springer Nature. Focused on leveraging LLM agents and NLP sequence modeling to build impactful, production-ready AI solutions.

SKILLS

Machine Learning & Generative AI: LLMs, RAG, AI Agents, Transformers, BiLSTM, Encoder-Decoder Architectures, Supervised/Unsu- pervised ML, Sentiment Analysis, Multimodal Fusion, Prompt Engineering, Speech-to-Text (STT), Text-to-Speech (TTS), GloVe, Word2Vec Programming Languages: Python, SQL, C, C++, HTML, JavaScript, Type Script ML Libraries & Tools: PyTorch, TensorFlow, Scikit-learn, NLTK, SpaCy, Langchain, HuggingFace, Node.js, FAISS, Chroma, Scikit learn, Gensim, Pandas, NumPy, Matplotlib, Gradio, Seaborn Frameworks & APIs: Flask, Streamlit, FastAPI, Gradio, PyTorch, TensorFlow, Django, Rest APIs, Node.js Professional Skills: Data Science, Software Development, Feature Engineering, Data Preprocessing, Model Validation, Data Analysis, Au- tomation, User Interaction

Cloud/ Deployment: Git, Docker, Firebase, Hugging Face Spaces, Cloudflare WORK EXPERIENCE

Maya IT Solutions

AI Engineer Jun 2025 - Present

• Built a RAG-based AI tutor bot (FAISS, embeddings, PDF retrieval), improving answer relevance (~30% ROUGE-L) and reducing query latency (~40%) via optimized chunking and retrieval strategies.

• Developed a AI agents based multimodal healthcare application integrating vision inputs and speech pipelines (LangChain, STT/TTS, FastAPI), enabling real-time inference and low-latency clinical query handling.

• Built and deployed Luv2Play conversational AI bot with dynamic dialogue flows, semantic FAQ matching, and confidence-tiered responses improving response consistency and reducing unanswered query rate by ~35%. Upwork

AI Engineer Apr 2026 - Present

• Optimized xLSTM based time series forecasting models for blood glucose prediction by improving preprocessing pipelines, feature engineering, and training strategies on large scale healthcare datasets.

• Performed feature selection, sample size optimization, and hyperparameter tuning to enhance model generalization and reduce computational overhead.

• Designed and executed experiments focused exclusively on xLSTM architecture optimization, achieving RMSE improvement from 25 to 19 through iterative tuning and data refinement.

• Worked with high volume sequential data (20M+ rows), implementing scalable data processing and efficient training workflows using Python and deep learning frameworks.

PROJECTS

Sarcasm Interpretation Model

• Constructed a context-aware sarcasm interpretation model using a BILSTM-based encoder-decoder with external word embeddings (SpaCy, GloVe, Word2Vec).

• Outperformed baseline model by 12.6% improvement in ROUGE score by addressing vanishing gradient issue. MCP-based Voice-Enabled Chatbot for Nutrition and Fitness

• Built and deployed an MCP-based voice AI chatbot delivering 24/7 personalized meal and workout plans to real users leveraging NLP, LLMs, and AI Agents for budget-aware, preference-driven recommendations

• Handles dynamic multi-turn conversations with context retention, generating fully customized plans based on individual user profiles and dietary constraints

EDUCATION

Kurukshetra University, India India

Ph.D. in Computer Science (Machine Learning, NLP) Oct 2019 - Jun 2025

• Thesis: “Review Based Analytics: Advancing Sentiment Analysis Using Artificial Intelligence” CERTIFICATIONS

• Deep Learning, Mastering Neural Networks: MIT xPRO, 2024

• AWS Cloud Bootcamp: ThinkCloudly, 2025

LANGUAGES

• Python: Primary • SQL

• C • C++

• HTML • JavaScript

• Type Script: Professional

PUBLICATIONS

Encoder-Decoder Model with Attention Mechanism for Sarcasm Interpretation IJIT, Springer Nature, 2025

Improving Sentiment Analysis Using Negation Scope Detection IJCDS, 2024

A Comparative Study of Supervised and Unsupervised ML Algorithms IEEE AIC, 2022

Contact this candidate