Hi, I'm Nageeta (Nigeeta) Kumari
AI/ML Engineer & Software Engineer with expertise in building intelligent systems using machine learning, natural language processing, and modern software engineering practices. I have completed my MVA at École Normale Supérieure Paris-Saclay and am currently exploring software engineering and ML engineering roles where I can apply my skills in developing scalable AI solutions and robust software systems.
Work Experience
Datadog – AI/ML Research Intern Apr 2025 – Oct 2025
❗ Problem: Distributed microservices produce complex trace data, making root cause analysis challenging and costly.
🧩 Strategy: Formulated the problem as a probabilistic modeling challenge, explored unbiased data sampling, and compared evaluation methods.
✅ Solution: Designed models for root cause analysis, implemented unbiased data pipelines, and built an evaluation framework achieving 95% recall & 77% precision.
🛠️ Tools: AWS, Trino-SQL, Docker, Mortor, NBT, Unsupervised Learning, Causal AI
📚 What I Learned: Applied AI research in production-scale systems, theory-engineering balance, multicultural teamwork.
INRAE – Research Intern Feb 2024 – Jul 2024
❗ Problem: INCA2 and INCA3 surveys had inconsistent nomenclature, blocking unified classification.
🧩 Strategy: Applied ontology mapping & semantic web tech to harmonize datasets.
✅ Solution: Built GPT-3.5 pipeline to map INCA2 into FoodEx2 for interoperability.
🛠️ Tools: GPT-3.5, Python, Ontology Dev, Semantic Web
📚 What I Learned: Data harmonization, semantic tech, handling real-world noisy data.
SoundM / DLLC – Software Engineer Aug 2022 – Aug 2023
❗ Problem: Need for secure, scalable communication platform.
🧩 Strategy: Designed secure back-end systems integrated with front-end features.
✅ Solution: Built chat systems, authentication, profile APIs, and deployed on AWS.
🛠️ Tools: React.js, Node.js, FastAPI, AWS
📚 What I Learned: Real-time systems, startup environment, cross-functional teamwork.
Freelancer – Fiverr 2020 – 2022
Fiverr Profile: fiverr.com/nageeta_w
❗ Problem: Clients needed secure web apps with fast delivery & low cost.
🧩 Strategy: Defined scope, built clean UI, deployed scalable solutions.
✅ Solution: Delivered full-stack projects (auth, APIs, dashboards) with high satisfaction.
🛠️ Tools: React, Node, FastAPI, PostgreSQL, MongoDB, AWS, Docker
📚 What I Learned: End-to-end project delivery, client management, solo development.
Education
École Normale Supérieure Paris-Saclay – MVA Sep 2024 – Sep 2025
Grade: 16/20
Selective AI Master's in computer vision, learning, and mathematics.
Courses:
- Reinforcement Learning
- Probabilistic Graphical Models
- Advanced Deep Learning
- Large Language Models
- Geometric Data Analysis
- Machine Intelligence of Images
Université Paris-Saclay – Master in Data Science Aug 2023 – Aug 2024
Grade: 15.7/20
Courses:
- Advanced Databases, Distributed Systems
- Deep Learning, NLP, ML Algorithms
- Information Theory, Probability
Sukkur IBA University – BS in CS Aug 2018 – Aug 2022
Grade: 3.68 / 4
🏅 Silver Medalist
Courses: Software Engineering, Object Oriented Programming, Web Enterprise Development, Database Design, Data Structures, Advanced Algorithms, Business Communication, Cloud and Parallel Computing, Robotics, Cyber Security, IoT
Projects
Modeling for MNAR Data (not-MIWAE)
Re-implemented not-MIWAE generative model and benchmarked it on UCI & stock datasets.
Tools: PyTorch, Python
Composed Image Retrieval
Adapted CoVR-BLIP-2 for CIRR dataset and improved Recall@1 with attention pooling.
Tools: Vision-Language Models, PyTorch
Patent Match Challenge
Used sentence-transformers and fine-tuned BERT models for patent citation matching.
Tools: NLP, Transformers
Compositional Understanding in VLMs
Evaluated VLMs on ARO benchmark, showing improved performance of Qwen2.5 over CLIP/BLIP.
Tools: Vision-Language Models, PyTorch
RelevAI-Reviewer (Benchmark System)
Relevance ranking for survey papers vs CFP prompts using ML baselines & fine-tuned transformers.
Tools: Transformers, Python
Scrimish (Android Game)
Android implementation of Scrimish card game with game logic and UI design.
Tools: Android Studio, Java
SIBA Stack Overflow (Q&A Web App)
Student/teacher Q&A platform with React.js, Node.js, MySQL.
Tools: React.js, Node.js, MySQL
Gender Detection Using Eye
CNN model predicting gender from eye images with preprocessing and training pipeline.
Tools: CNN, Python, TensorFlow
Publications
RelevAI-Reviewer: A Benchmark on AI Reviewers for Survey Paper Relevance
Authors: Paulo Henrique Couto, Quang Phuoc Ho, Nageeta Kumari, Benedictus Kent Rachmat, Thanh Gia Hieu Khuong, Ihsan Ullah, Lisheng Sun-Hosoya
Conference: Conférence sur l'Apprentissage Automatique (CAp) 2024, Lille, France
Abstract: Recent advancements in AI, particularly Large Language Models (LLMs), offer promise for automating scientific paper review. This paper proposes RelevAI-Reviewer, an automatic system that conceptualizes survey paper review as a classification problem, aimed at assessing paper relevance to specified prompts. We introduce a novel dataset of 25,164 instances and explore various baseline approaches including traditional ML classifiers like SVM and advanced language models such as BERT.
Awards & Honors
Silver Medal Apr 2023
Issued by: Sukkur IBA University
Silver Medalist for achieving the second-highest position based on Bachelor's CGPA in Batch 2018.
Robotics Competition May 2022
Issued by: CRAIB Lab, Sukkur IBA University
Participated in a robotics competition where students assembled and programmed robots for specific tasks. Worked with a team on a spider robot using 3D printed parts, programmed to access hidden areas that are hard for larger robots to reach.
Merit Scholarship 2019 Dec 2019
Issued by: Sukkur IBA University
Institutional merit-based scholarship awarded to the Top 3 students of the department based on Fall semester GPA for the entire academic year.
Mathematics Olympiad Winner Aug 2018
Issued by: Sukkur IBA University
Won the Mathematics Olympiad competition hosted by Sukkur IBA University with its sub-campuses for foundation semester Batch 2018.
Contact Me
I'd love to hear from you! Feel free to reach out via email or connect with me on LinkedIn.
Email: nageetaw@gmail.com
LinkedIn: linkedin.com/in/nageeta124
GitHub: github.com/nageetaw