Rakshita Iyer

AI Engineer | Python, ML, NLP, LLMs

AI Engineer with strong expertise in Python, statistical modelling, and predictive analytics. Experienced in applying machine learning to real-world datasets across education, healthcare, and social media. Skilled in regression, clustering, and classification using pandas, scikit-learn, and SQL, with hands-on work in TensorFlow, PyTorch, and modern ML libraries.

About

Structured problem-solving, real-world ML, and impact-focused AI.

I am an AI Engineer with a strong foundation in Python, machine learning, and deep learning. I enjoy building end-to-end systems — from data preprocessing and modelling to evaluation and deployment , with a focus on reliable, production-ready solutions. My work spans student behaviour analysis, healthcare applications, and social media analytics, always guided by first-principles thinking and data-driven decision making.

Skills

Technologies and tools I work with.

Programming

Python, SQL, MATLAB, C++

ML & AI

TensorFlow, PyTorch, Keras, scikit-learn, NLP, LLMs, BERT, GPT-based models, clustering, regression, classification, statistical modelling.

Tools & MLOps

NumPy, Pandas, OpenCV, NLTK, FastAPI, Flask, Docker, Git, Weights & Biases, CI/CD, FAISS.

Data & Analytics

A/B testing, hypothesis testing, data visualisation, SQL-based analysis.

Projects

Selected work in AI, NLP, and healthcare.

Hackathon Winner • London • October 2024

Multi-Agent Cryptocurrency Trend Prediction System

Built an end-to-end multi-agent system to analyze social media data and forecast cryptocurrency trends, integrating live Twitter feeds and ML models for sentiment and trend detection.

Impact: Processed and analysed 10,000+ tweets daily, improving sentiment classification precision by 18% and reducing API response time by 35% using Docker and uAgents for scalable deployment.

Tech: Python, TensorFlow, OpenCV, Twitter API, AI Agents, Docker

Hackathon Winner • London • April 2025

MediSync: Decentralized Healthcare Platform

Designed a decentralized, offline-first healthcare platform with interoperable apps (PatientConnect, DoctorDesk, LabLink) to enable secure communication and medical data sharing without centralized databases.

Impact: Integrated AI-powered vitals analysis, MRI tumor detection, and medical image segmentation via custom UNet, delivering real-time health insights, tumor classification, and lesion segmentation masks.

Tech: Python, TensorFlow, PyTorch, Flask APIs, HuggingFace Transformers, JS

NLP Project • March 2024 – June 2024

Sequence Labelling for Imbalanced NLP Dataset

Developed and fine-tuned a Bi-LSTM model for sequence labelling on a highly imbalanced dataset, comparing multiple baselines and encoding strategies.

Impact: Improved F1-score by 18%, reduced training time by 15% through careful hyperparameter tuning, and increased recall for minority entities by 25% using oversampling techniques.

Tech: Python, NLTK, Hugging Face, Seaborn, Bi-LSTM, Word2Vec

Experience

Professional background and responsibilities.

Associate System Engineer • Tata Consultancy Services (TCS)

June 2020 – July 2021 • Mumbai, India

Collaborated with cross-functional international teams in Agile environments to ensure software reliability.
Conducted manual testing and QA, identifying performance bottlenecks and improving system stability.
Coordinated testing pipelines and presented findings in weekly case review meetings.
Developed strong foundations in system scalability, CI/CD workflows, metrics tracking, and data-driven problem solving.

Education

Academic qualifications.

M.Sc. in Artificial Intelligence • University of Surrey

Feb 2024 – Feb 2025 • England

Coursework: Deep Learning, NLP, Computer Vision, Machine Learning, AI Ethics, Reinforcement Learning.

B.E. in Computer Engineering • SIES Graduate School of Technology

Aug 2017 – Aug 2021 • India

Coursework: SQL, Machine Learning, Python, C++, Data Structures & Algorithms, Graph Theory, Algorithms.