Aerogeophysical ETL & ML Pipeline
Python · Dask · XGBoost · Streamlit · PyDeck
145GB geophysical survey processing pipeline across 19.3M points. Includes FFT-based feature extraction and 5-page geospatial UI.
data scientist · ai engineer
turning complex problems into systems that actually work
Hi, I'm Chandan.
I'm an AI Engineer in the making, fascinated by the idea of turning complex problems into systems that actually work.
My journey started with curiosity about why user data is stored and maintained, and what makes predictions useful. That curiosity gradually turned into a habit of building — experimenting with models, developing applications, working with data, and constantly trying to understand what happens beneath the surface.
I work across Artificial Intelligence, Machine Learning, Data Science, and Software Engineering, with a particular interest in building practical AI systems rather than keeping ideas confined to notebooks.
I've worked on projects involving LLMs, explainable AI, data analytics, privacy-preserving AI, machine learning, full-stack applications, and intelligent automation. One of the things I enjoy most is taking an idea that initially feels complicated or even unrealistic and breaking it down until I can build a working version of it.
I want to understand how AI models work, how they fail, and how they can be engineered into reliable systems. My projects are where I learn. Every project gives me an opportunity to explore a new technology, solve a new problem, make mistakes, and push my understanding a little further.
I'm particularly interested in the intersection of AI and real-world engineering — where models meet data, software, users, constraints, and actual problems. I enjoy working with technologies such as Python, PyTorch, NumPy, Pandas, SQL, MongoDB, React, Flask, FastAPI, and modern AI/LLM frameworks, while continuously expanding my technical stack.
Python · Dask · XGBoost · Streamlit · PyDeck
145GB geophysical survey processing pipeline across 19.3M points. Includes FFT-based feature extraction and 5-page geospatial UI.
Kafka · Spark Streaming · HDFS · TimescaleDB
Live Binance WebSocket ingestion with Spark Structured Streaming, HDFS Parquet lakehouse, regime classification, and Redis state.
XGBoost · DistilBERT · SHAP · CLIP · spaCy
Ensemble ML system for identifying coordinated inauthentic online campaigns across multiple sources with explainability layers.
FastAPI · RAG · Groq API · SQLite · Railway
Emergency response bot delivering certified protocols via online RAG and offline SQLite retrieval, engineered for critical voice/text I/O.
PyTorch · OpenCV · Deep Learning
Specialized vision model addressing localized vehicular classification under heterogeneous Indian traffic dynamics and illumination.
Python · Scapy · Network Protocols
Automated network diagnostic and packet auditing toolkit designed for controlled wireless security assessments and latency discovery.
Computer Vision · Food Recognition · Analytics
Nutritional estimation engine mapping visual meal captures to macronutrient decompositions with instant health metric logs.
React · TypeScript · University Schema
Full-stack academic utility tailored to VTU grading schemes, weighted credit distribution algorithms, and forecast modeling.
Vision Transformers · NLP · PyTorch
Automated e-commerce catalog descriptor generating descriptive contextual captions using cross-attention vision-language backbones.
IBM SkillsBuild
Data governance, exploratory analysis, and analytical workflows.
Krish Naik — Udemy
Core mathematical ML, algorithmic implementations, and model evaluation.
Krish Naik — Udemy
Distributed computation patterns, Hadoop HDFS ecosystems, and MapReduce.
Genesis
Business intelligence pipelines, DAX computations, and interactive executive reporting.
Great Learning Academy
Object-oriented paradigms, concurrency frameworks, and data structures.
Demonstrated high-pressure engineering capability across competitive hackathons, prototyping functional end-to-end applications within 24 to 36-hour sprint windows.
Sambhram Institute of Technology
I'm a motivated engineering student who is naturally curious and enjoys exploring new ideas, technologies, and tools.
Outside my core AI and data-science work, I enjoy experimenting with cybersecurity and technical tools in controlled learning environments. I've explored tools such as Termux, John the Ripper, and Kali Linux utilities as part of learning about security concepts, systems, and how different techniques work.
I also enjoy learning about social engineering from a security-awareness perspective, particularly understanding how human behavior can become part of a security system's attack surface.
I read occasionally, go on treks when I get the opportunity, and enjoy exploring things that challenge the way I think.
One thing I value about myself is persistence. Once I start something, I find it difficult to leave it unfinished — even when it is something small. I like reaching the point where I can finally say, "done."
That persistence is probably one of the biggest reasons I keep building.
Curious by default.
Building by choice.
Still figuring out what's next.