SEKH ASIF SOHAL

Bachelor of Computer Application student focused on artificial intelligence, machine learning, and data science

New Delhi, IndiaArtificial Intelligence and Machine Learning
4roles
39skills
2education
4credentials

About

Computer application student with hands-on experience building AI, machine learning, and Agentic RAG projects using Python and modern ML frameworks.

Experience

Artificial Intelligence Intern

Codec Technologies

Apr 2026 – May 2026 · Hybrid, India

Executed end-to-end machine learning workflows including data cleaning, Min-Max scaling, and feature engineering; optimized model performance using GridSearchCV and k-fold cross-validation across CNN-based image recognition and NLP pipelines.

Core Member

Cyber Hub Club, DSEU

Orchestrated campus technical events and promotional design.

Participant

Industrial Ideathon 2025

2025

Pitched technical solutions for urban air pollution monitoring.

Photography Team Head

Khabre Vidhyarthi Media

Led event coverage and social media curation.

Education

Delhi Skill and Entrepreneurship University, Ambedkar Campus

Bachelor in Computer Application

Oct 2023 – 2026

Rajkiya Sarvodaya Bal Vidyalaya, Gandhi Nagar

CBSE, Science & Computer Science

May 2023

Skills

PythonSQLC++JavaBash/Linux ShellMachine LearningDeep LearningGenerative AILLMsAgentic RAGAI AgentsData EngineeringFeature EngineeringLinear AlgebraCalculusProbabilityPyTorchTensorFlowScikit-LearnLangChainLangGraphHugging FacePandasNumPySupabase pgvectorPineconeChromaDockerKubernetesFastAPIFlaskRenderAWSAzureGitGitHubMLflowPrompt EngineeringDebugging Model Hallucinations

Projects

Agentic RAG Web Assistant

Engineered a full-stack Agentic RAG application with Python and FastAPI, featuring a LangGraph AI agent that routes queries between live web searches and internal document retrieval. Used Supabase PostgreSQL with pgvector for semantic search of embedded PDFs and Docker for cloud hosting on Render. Technologies: Python, FastAPI, LangGraph, PostgreSQL, Docker. Date: June 2026.

Handwritten Digit Recognizer

Developed a four-layer CNN in TensorFlow/Keras achieving 99.2% accuracy on MNIST, with a real-time Streamlit drawing canvas and Plotly data visualizations. Technologies: Python, TensorFlow, Streamlit, Plotly. Date: May 2026.