Hi, I'm Athar Sayed

Agentic AI Engineer

View Projects Download Resume
0
Years Experience
0
Projects
0
Technologies

About Me

Portrait of Athar Sayed smiling

Hi, I'm Athar Sayed, an AI-ML engineer who builds production ready systems that deliver measurable impact.

I am pursuing an M.Tech in Artificial Intelligence at NMIMS and working as a Gen-AI Intern at Diebold Nixdorf. I focus on building scalable GenAI systems, agentic workflows, RAG pipelines, infrastructure, and reliable deployment.

Interested in collaborating on applied AI or privacy-first data systems? Let's connect.

M.Tech — Artificial Intelligence

Mukesh Patel School of Technology Management & Engineering 2024 — 2026 CGPA: 9.10 / 10

B.Tech — Electronics & Telecommunication

Symbiosis Institute of Technology 2019 — 2023 CGPA: 7.11 / 10
LeetCode LeetCode HackerRank HackerRank

Featured Projects

Der Kurator RAG System

Der Kurator: Document Grounded RAG System

  • Built an end-to-end document-grounded RAG system covering ingestion, retrieval, generation, and evaluation.
  • Implemented multi-format ingestion (PDF, DOCX, PPTX, TXT) with structure-aware chunking to preserve numeric and variant-specific data.
  • Developed FAISS-based semantic retrieval with variant-aware filtering, diversity control, and safe fallback logic.
  • Added multi-layer hallucination control with strict grounding prompts, answer gating, citations, and safe abstention.
  • Results: 97% grounded sentence rate, 99% faithfulness, and controlled hallucination behavior.
Code
Vigilix

Vigilix: Network Intrusion Detection System

  • Developed a Network Intrusion Detection System (NIDS) using the UNSW-NB15 dataset, achieving 80%+ accuracy in intrusion detection.
  • Evaluated ML models (XGBoost, Random Forest, Isolation Forest), with XGBoost achieving 87.95% accuracy, 85.35% precision, and 89.61% F1-score.
  • Optimized XGBoost performance through hyperparameter tuning, improving accuracy by 13% and F1-score by 5.25%.
  • Automated end-to-end workflow with Kafka, Prometheus, Grafana, and a custom dashboard, enabling 99.9% uptime.
Live Demo Code
IntelliTube

IntelliTube: AI Powered YouTube Insight Engine

  • Developed end-to-end YouTube video analysis platform with 100% local execution
  • Achieved 95% transcription accuracy using Faster Whisper
  • Multilingual sentiment analysis with 89% accuracy
  • Secure user authentication for 50+ concurrent users
  • Question-Answering module leveraging FAISS vector search, MPNet embeddings, and an Ollama-powered Mistral LLM, achieving ~35% faster retrieval, 20–25% improved answer relevance, and significantly reduced hallucinations through recursive text-chunking and a grounded QA prompt.
Live Demo Code
ScriptSense

ScriptSense: Personality Prediction

  • Developed ScriptSense, a hybrid CNN-graphology tool predicting Big Five personality traits from handwriting images using TensorFlow/Keras and OpenCV for feature extraction.
  • Integrated CNN predictions with graphology rules to generate combined, confidence-scored trait insights.
  • Predicts introversion/extroversion, emotional stability
  • Deployed on Hugging Face Spaces
Try It Code
Academic Dashboard

Academic Dashboard Mtech Students

  • Collected and preprocessed academic and demographic data for 44 M.Tech students, ensuring 100% data accuracy through thorough cleaning and validation techniques.
  • Analyzed data to extract key insights, leading to improved visualization strategies and highlighting trends in academic performance and participation.
  • Designed and implemented over 10 interactive visualizations using tools like Matplotlib and Plotly to represent student metrics on a user-friendly dashboard.
  • Deployed the dashboard on Streamlit Cloud, achieving seamless access and reducing manual reporting efforts by over 80%.
View Code
StreamPulse Dashboard

StreamPulse : Real-Time E-Commerce Transactions Dashboard

  • Built real-time e-commerce transaction monitoring system using Kafka + Spark Structured Streaming + Streamlit, processing 1,000+ events/min with <2s latency.
  • Implemented end-to-end streaming pipeline: Kafka producer simulating e-commerce events → Spark job for data cleansing and aggregation → live-updating Plotly dashboard with dark theme.
  • Built 8+ interactive visualizations showing real-time revenue, transactions per minute, top products, Total Sales by country , and customer geographic distribution using PySpark and Streamlit.
View Code

Blogs

Recent posts on Medium — click any card to read the full article.

Space Data Centers — The Next Frontier of Data Computation

Medium · Recent

A short primer on why placing data centers in space could change latency, resilience, and the economics of large-scale computation.

Read the full article

More on Medium

Visit my profile

Find more articles and technical write-ups on my Medium profile. Click to view all posts and follow for updates.

Visit Medium Profile

Publications

IEEE ICRTEC 2023

IoT Based Crowd Detection and Stampede Avoidance using Predictive Analysis

Authors: Athar Sayed, Harikrishnan R., et al.

Developed an IoT-powered crowd monitoring system with real-time predictive analytics. Leveraged computer vision and ML algorithms to detect high-risk stampede conditions with 92% accuracy, improving large-event safety outcomes.

IEEE Xplore View PDF

Achievements

10+

AI Projects Completed

2

Research Papers Published

5

Pull Requests Merged

900+

GitHub Contributions (2025)

150+

LeetCode Problems Solved

Technical Skills

Programming & Analytics

  • Python
  • C++, C
  • DSA

AI & Machine Learning

  • TensorFlow, PyTorch
  • OpenCV, Keras
  • YOLO (Object Detection)

DevOps & MlOps

  • Docker, Kubernetes
  • Git, GitLab, CI/CD
  • GitHub Actions,Jenkins
  • Prometheus, Grafana

LLM & Agentic AI Frameworks

  • LangChain
  • RAG (Retrieval-Augmented Generation)
  • Agentic AI Workflows
  • Model Context Protocol (MCP)

Databases

  • MongoDB
  • PostgreSQL
  • MySQL

Work Experience

Gen-AI Intern @ Diebold Nixdorf

February 2026 — Present
  • Contributing to enterprise-scale Generative AI solutions by enhancing an internal LLM-powered chatbot through prompt engineering, response evaluation, and output optimization.
  • Optimizing the Retrieval-Augmented Generation (RAG) pipeline by refactoring preprocessing modules, cleaning legacy code, and restructuring functions for improved modularity and maintainability.
  • Conducting prompt experimentation and LLM output evaluation to analyze model behavior, improve contextual responses, and reduce hallucinations.
  • Improving LLM workflow efficiency by optimizing context handling and ensuring response consistency across the chatbot system.

Software Engineer @ FortytwoLabs

June 2023 — June 2024 (1 year 1 month)
  • Built and maintained 6+ production-ready C++ libraries with cross-platform support (Linux, Windows, macOS), achieving 100% build stability and reducing integration issues by 30%
  • Implemented cryptographic features (e.g., policy signature modules), enhancing security and reducing processing time by 40%.
  • Created 143 unit tests using Google Test (GTest) and JUnit, achieving 95% code coverage and reducing bugs by 50%.
  • Used Postman and JMeter for API and performance testing, improving system performance by 25%.
  • Documented 100+ error codes, improving client troubleshooting efficiency and decreasing support tickets by 40%.
  • Collaborated with QA, DevOps, and Product teams, contributing to faster and more reliable software releases.

Technical Intern @ FortytwoLabs

January 2023 — May 2023 (5 months)
  • Contributed to maintaining CI/CD pipelines by writing YAML scripts and debugging build issues under guidance.
  • Assisted in configuring Nginx and Linux-based reverse proxy setups to optimize application routing & improve deployment reliability
  • Configured Apache ActiveMQ within Docker containers on CentOS, enabling reliable message brokering and client-server communication.
  • Deployed Java-based web applications by packaging and deploying WAR files on Apache Tomcat, ensuring stable server-side execution.

Certifications

LLM, Vertex AI, Gemini & Principles of Google Cloud

Google Cloud

Complete Python from Beginner to Master

Udemy

Intermediate Machine Learning

Kaggle

Linux Mastery Mastering the Command Line

Udemy

Mastering DSA using C and C++

Udemy

Oracle Logo

Oracle Cloud Infrastructure 2025 Certified AI Foundations Associate

Oracle

Oracle Logo

Oracle Cloud Infrastructure 2025 Certified Generative AI Professional

Oracle

IRIS Intelligent Responsive Interface System
Llama 3.1