Who Am I?
Jashwanth S.
Turning raw data into real-world AI solutions.
I'm an AI/ML Engineer specializing in deploying robust machine learning systems. Currently, I'm focused on building accessible, human-centered products and exploring the world of Generative AI.
E-City, Bangalore | jashwanthsangu07@gmail.com | +91-9019702498
01. Where Did I Learn All This?
Visvesvaraya Technological University
- BE (Honors) in Artificial Intelligence & Machine Learning
- CGPA: 9.53/10.0
- Bangalore, Karnataka
02. Where I Put Skills Into Action and Excelled!!!
AI Engineer @ GlobalData PLC
Aug 2025 - Present
- Built LLM-powered automated tagging workflows integrated with Spark, combining NER, TF-IDF/KeyBERT keyword extraction, and embedding-based semantic search to tag content at scale; used zero/few-shot LLM prompting via API for context-aware classification, reducing manual tagging effort by 70%.
- Automated data pipelines with Apache Airflow, orchestrating multi-step DAGs for ingestion, transformation, and validation across 1M+ rows weekly; incorporated retry logic, alerting, and dependency management to ensure fault-tolerant execution, cutting manual effort by 60%.
- Developed Spark-based enrichment pipelines handling 100GB+ datasets, applying distributed joins and custom UDFs to enrich records against internally validated mapping files and LLM-generated annotations, ensuring consistent, high-quality feature enrichment at scale.
- Built a data collection pipeline harvesting Google and YouTube search data using Playwright for browser automation with header caching and intelligent request throttling to avoid API rate limits; integrated LLM-based content classification and relevance scoring, with data stored in S3 and PostgreSQL, improving ingestion speed by 40%.
- Built and automated classification and prediction jobs for 10M+ data points across 200+ dashboards, incorporating metrics like CAGR and rolling windows with MCP integrations, triggered monthly with intelligent monitoring and retry mechanisms.
Data Engineering Intern @ Ai Palette
Aug 2024 - Apr 2025
- Managed 10M+ records across production datasets, ensuring high availability and reducing data retrieval errors by 25%.
- Engineered optimized real-time processing workflows, increasing system throughput by 35% and improving response times by 20%.
- Built scalable pipelines automating data movement to production, cutting manual processing time by 40% and enabling near real-time analytics.
03. Built With Curiosity + Coffee
Predictive Maintenance for IoT
Engineered an ML system using Python, Scikit-learn, and TensorFlow to forecast IoT device failures, achieving 92% accuracy.
Chat with my Data (RAG System)
Engineered a Retrieval-Augmented Generation (RAG) system to process 5,000+ documents, achieving 2.3-second average retrieval time.
Waste Management System
Engineered a CNN-based waste detection system that classified 10 distinct material categories with 92% test accuracy.
04. Everything in My Arsenal
Programming Languages
- Python
- Java
- SQL
- R
- JavaScript
- HTML/CSS
AI/ML Frameworks & Libraries
- TensorFlow
- PyTorch
- Scikit-learn
- XGBoost
- Pandas & NumPy
- LangChain & Transformers
- Ollama
- FastAPI & Streamlit
Databases & Data Stores
- MySQL
- MongoDB
- Elasticsearch
- Vector Databases
Platforms & Tools
- AWS
- Google Cloud Platform
- Docker
- Kubernetes
- Git & GitHub
- Jupyter Notebooks
Core Concepts & Specializations
- Deep Learning & NLP
- Computer Vision
- Generative AI & RAG
- LLM Fine-tuning
- AI Agents
- MLOps
- Predictive Modeling
- Feature Engineering
- Statistical Analysis
- Data Structures & Algorithms
05. The Story Behind the Engineer
What Drives Me
"I believe AI should be useful, reliable, and accessible. My focus is building systems that combine strong engineering with real user impact—not just model accuracy."
Breakthrough & Growth GlobalData PLC
Integration of LLMs into ETL systems and development of sentiment models optimized for production environments.
"Earned Rewards & Recognition Awards for consistently transforming learning into execution."
Stepping into Industry
Engineered scalable data pipelines processing over 10M+ records. Improved throughput by 35% and reduced manual workload by 40%.
Learned that AI isn’t just about architecture—it’s about respecting the data, understanding its behavior, and designing pipelines that make models truly intelligent.
Building Real-World Solutions
Transitioned from academic models to solving real problems—like forecasting failures in IoT devices and building efficient RAG systems that reduced query latency by 2.3 seconds for large datasets.
Key Realization: Even advanced models collapse without the right data foundation.
From Curiosity to Craft
It started with a simple question: "How do machines learn?"
Built my first handwritten digit classifier. Watching a model learn from data fascinated me and sparked a drive to explore intelligence at scale.
Vision Ahead
Currently exploring AI Agents, low-cost inference, and real-time ML systems—aiming to build products that scale to millions, not just benchmarks.
"My long-term goal is to build scalable and powerful AI systems that solve real-world problems at meaningful scale — systems that don’t just work in research papers, but in everyday life."
06. Let’s Talk
Get In Touch
I'm currently open to new opportunities and collaborations. Whether you have a question or just want to say hi, I’ll try my best to get back to you!
Say Hello