Argish Abhangi's Resume

Argish Abhangi

Detail-oriented Machine Learning Engineer dedicated to building high-quality AI Powered products.

Surat, India

About

Machine Learning Engineer with a strong foundation in Machine Learning and Deep Learning. While I am just starting my professional journey, I am passionate about leveraging cutting-edge algorithms and techniques to solve complex problems and drive innovation. My goal is to contribute to impactful projects while continuously enhancing my skills in this dynamic and ever-evolving field.

Work Experience

Spigot InfoTech

March 2024 - May 2024

AI/ML Intern

  • Developed and deployed a web scraping pipeline leveraging BeautifulSoup & Selenium to auto-scrape data from 30+ websites, boosting data extraction efficiency by 80% & reducing manual effort by 50 hours monthly.
  • Designed RAG pipelines and engineered prompt optimization techniques to improve response quality.
  • Collaborated with cross-functional teams to integrate AI-driven chatbots and summarization tools.
  • Langchain
  • OpenAI
  • Ollama
  • CrewAI
  • Streamlit

Space Applications Centre - ISRO

June 2024 - July 2024

Machine Learning Intern

  • Developed a novel deep neural network architecture to automate the alignment of primary and secondary mirror segments for a telescope, enhancing precision and reducing manual intervention.
  • Utilized residual attention networks (RANs) & multi-scale feature extraction to optimize image processing tasks.
  • Proposed real-time feedback mechanisms to dynamically adjust mirror positioning for optimal alignment.
  • Collaborated with astrophysicists & optical engineers to understand telescope architecture & requirements, gather domain knowledge and integrate AI-driven solutions into the telescope's control systems.
  • Python
  • PyTorch
  • OpenCV
  • Zemax

Water Management and Forecasting, Omdena

-

Open Source

  • Developed, implemented & fine-tuned LSTM Network using recursive forecasting to predict water availability with 96% accuracy using sensor data from the cities of Algiers & Bhopal. Deployed the model using Streamlit.
  • Python
  • PyTorch
  • Scikit-Learn
  • Plotly
  • Streamlit

Greenhouse Monitoring Systems, Omdena

-

Open Source

  • Built a plant disease & pests detection CNN model to identify 40+ diseases & pests with 98% accuracy.
  • Fine-tuned a plant species identification model using transfer learning to identify 20 species with 94% accuracy.
  • Python
  • Tensorflow
  • Keras
  • OpenCV
  • HTML
  • CSS
  • JavaScript

Education

Vellore Institute of Technology, Bhopal

2020 - 2025
M.Tech AI (5 Year Integrated) | CGPA: 8.89

Ryan International School, Surat

2019 - 2020
Secondary Higher Education (XII) | Grade - 85%

Ryan International School, Surat

2017 - 2018
Higher Education (X) | Grade - 73%

Skills

  • Python
  • C++
  • SQL
  • PyTorch
  • LangChain
  • LangGraph
  • CrewAI
  • HuggingFace Transformers
  • Tensorflow
  • Keras
  • NLTK
  • OpenCV
  • FastAI
  • ChromaDB
  • SQLite
  • Numpy
  • Pandas
  • FireDucks
  • Scikit-learn
  • XGBoost
  • BeautifulSoup
  • Selenium
  • FastAPI
  • Flask
  • Streamlit
  • Matplotlib
  • Seaborn
  • Plotly
  • HTML/CSS
  • Git/Github

Side projects

Realtime Multimodal Emotion Recognition | 🛠️ In Progress

Training a Text + Audio + Vision multimodal for state-of-the-art realtime emotion recognition.

  • Python
  • PyTorch
  • Tensorflow
  • Keras
  • Huggingface-Transformers
  • KaggleHub
  • Mediapipe
  • PyQT5

ChatDocs

Developed a robust RAG-based chatbot with LangChain & OpenAI, achieving 95% context accuracy. Integrated the RAG pipeline with a Streamlit app via FastAPI & SQLite for efficient session management. Optimized system performance & user experience through intuitive architecture & accurate response delivery.

  • Python
  • Langchain
  • Openai
  • FastAPI
  • ChromaDB,
  • SQLite
  • Pydantic
  • Streamlit

StockGinie

Built a scalable stock price prediction system using LSTM, achieving an average R² score of 0.85. Automated data analysis, model training & inference with real-time data fetching & comprehensive logging. Deployed an interactive Streamlit dashboard for visualizing insights with dynamic graphs & charts.

  • Python
  • PyTorch
  • Yfinance
  • Plotly
  • PyTest
  • Streamlit

genfunc

Shipped a python package that leverages power of LLMs to generate functions based on a given description. It reduces time taken to explicitly code one-time use functions & is useful to generate complex boilerplate code.

  • Python
  • OpenAI
  • Anthropic
  • Logging
  • PyTest