Skip to content
View Eshal-ven's full-sized avatar

Block or report Eshal-ven

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Eshal-ven/README.md

About Me

I am an AI Engineering student and AI Architect focused on building production-oriented AI systems rather than tutorial projects or toy demos.

My work focuses on Retrieval-Augmented Generation, multimodal retrieval, LLM integration, voice AI, speech pipelines, vector search, and applied machine learning.

I care about the engineering underneath an AI product: retrieval quality, grounding, hallucination control, latency, orchestration, observability, deployment, and maintainable architecture.

I am currently building a focused AI engineering portfolio while pursuing opportunities in AI/ML engineering, RAG engineering, LLM engineering, and applied AI development.

Open To

AI Engineer · ML Engineer · RAG Engineer · LLM Engineer · AI/ML Internships · Freelance AI Engineering · Remote Opportunities

Tech Stack

Languages

Additional: SQL · 8086 Assembly

AI / Machine Learning

  • LangChain
  • Gemini API
  • Groq
  • Whisper
  • ElevenLabs
  • Hugging Face
  • Scikit-learn
  • Pandas
  • NumPy

Web & Deployment

  • Streamlit
  • Streamlit Cloud

Development & Tooling

AI / ML Expertise

Domain Proficiency Details
Retrieval-Augmented Generation Advanced Multimodal RAG, hybrid retrieval, grounding and retrieval pipelines
Multimodal AI Advanced Text, image and table retrieval
LLM Engineering Advanced Gemini, Groq and LangChain integration
Vector Search Advanced Dense embeddings and semantic retrieval
Reranking Advanced Cross-encoder reranking and relevance optimization
Voice AI Advanced STT, TTS, VAD and interruption handling
Speech AI Advanced Groq Whisper and ElevenLabs
Hallucination Control Advanced Retrieval-grounded response generation
Prompt Engineering Advanced Structured prompts and LLM behavior control
Machine Learning Intermediate Scikit-learn and applied machine learning
Data Analysis Intermediate Pandas, NumPy and applied statistics
Conversational AI Advanced Rule-based and LLM-powered chatbot systems

Featured Projects

🎙️ CrossIndexRAG Voice Agent

A voice-interfaced conversational AI system built on top of CrossIndexRAG, allowing users to communicate with a multimodal retrieval system through natural speech.

The pipeline processes:

Voice Input → STT → Retrieval / Reasoning → Grounded Response → TTS

Engineering Highlights

  • Groq Whisper for speech-to-text
  • Gemini for reasoning
  • ElevenLabs for text-to-speech
  • Live microphone input
  • Custom energy-threshold Voice Activity Detection
  • Barge-in handling for natural interruption
  • Per-stage latency instrumentation
  • Grounded responses through CrossIndexRAG
Metric Implementation
Stack Python · Groq Whisper · Gemini · ElevenLabs · RAG
Architecture Voice interface over multimodal RAG
Performance Per-stage latency instrumentation
Reliability Custom VAD and interruption handling
Impact Voice-first interaction with a grounded RAG system

Repository: https://github.com/Eshal-ven/crossindexrag-voice-agent


🔍 CrossIndexRAG

An open-source multimodal hybrid Retrieval-Augmented Generation system designed to retrieve and combine information across multiple modalities before generating grounded responses.

Engineering Highlights

  • Dense embeddings per modality
  • Text retrieval
  • Image retrieval
  • Table retrieval
  • Cross-encoder reranking
  • Weighted score fusion
  • Hallucination control
  • Multimodal retrieval architecture
Metric Implementation
Stack Python · Embeddings · Vector Search · Reranking
Architecture Multimodal hybrid RAG
Retrieval Dense retrieval + reranking + score fusion
Reliability Hallucination control layer
Impact Open-source multimodal RAG foundation

Repository: https://github.com/Eshal-ven/CrossIndexRAG


🏠 Real Estate ROI Analyzer

A deployed Streamlit application that transforms property information into investment-oriented ROI analysis.

Engineering Highlights

  • Property ROI calculations
  • Interactive analysis
  • Data-driven investment metrics
  • Streamlit interface
  • Cloud deployment
Metric Implementation
Stack Python · Streamlit · Pandas · NumPy
Architecture Interactive web application
Deployment Streamlit Cloud
Impact Converts property data into investment metrics

Live Application: https://roi-analyzer-app.streamlit.app/


💬 Instagram Business Chatbot

An automated customer interaction chatbot developed for a small-business Instagram account.

Engineering Highlights

  • Automated customer interactions
  • Conversational workflows
  • Rule-based response handling
  • Business-focused chatbot architecture
Metric Implementation
Stack Python · Chatbot Architecture
Architecture Rule-based conversational system
Impact Automates repetitive customer communication

Experience

AI Researcher — Swan Labs

LLM Integration & RAG System Development

Worked on AI engineering tasks involving LLM integration and Retrieval-Augmented Generation systems.

Scope of Work

  • LLM integration
  • RAG system development
  • Retrieval pipeline development
  • AI system experimentation
  • Generative AI engineering

Skills: Python RAG LLMs LangChain AI Engineering

Python Developer Intern — Arch Technologies

Python Development & Chatbot Engineering

Worked on Python development, rule-based conversational systems and technical project implementation.

Scope of Work

  • Built and deployed a rule-based chatbot
  • Developed Python applications
  • Built a Dice Rolling Simulator
  • Built a To-Do List application
  • Prepared formatted technical project reports

Skills: Python Chatbot Development Software Engineering Git

Achievements

Recognition Details
Open Source Author Authored and open-sourced CrossIndexRAG, a multimodal hybrid RAG system
AI Systems Engineering Authored and open-sourced CrossIndexRAG Voice Agent
Production AI Focus Building practical RAG and voice AI systems beyond tutorial-level implementations
Deployment Deployed the Real Estate ROI Analyzer through Streamlit Cloud

Open Source

CrossIndexRAG

Multimodal hybrid RAG system.

https://github.com/Eshal-ven/CrossIndexRAG

CrossIndexRAG Voice Agent

Voice-interfaced AI system built on top of CrossIndexRAG.

https://github.com/Eshal-ven/crossindexrag-voice-agent

GitHub Analytics


GitHub Trophies

Contribution Activity

Contribution Snake

GitHub Contribution Snake

Current Focus

learning:
  - Advanced RAG architectures
  - Multimodal retrieval
  - LLM engineering
  - Voice AI
  - Production AI engineering

building:
  - CrossIndexRAG
  - CrossIndexRAG Voice Agent
  - AI engineering portfolio projects
  - Production-oriented AI applications

exploring:
  - AI/ML engineering
  - Retrieval optimization
  - Voice-first conversational systems
  - Full-stack AI products

open_to:
  - AI Engineer roles
  - ML Engineer opportunities
  - AI/ML internships
  - Freelance AI engineering
  - Remote opportunities

Coding Profiles

Connect

“Build AI systems that go beyond demos.”

Popular repositories Loading

  1. real_estate_ai_tool real_estate_ai_tool Public

    An AI-powered real estate ROI analyzer built with Streamlit.

    Python

  2. Eshal-ven Eshal-ven Public

  3. CrossIndexRAG CrossIndexRAG Public

    Multimodal RAG system with hybrid retrieval, cross-encoder reranking and hallucination control

    Python

  4. crossindexrag-voice-agent crossindexrag-voice-agent Public

    Real-time AI voice agent combining multimodal hybrid RAG, cross-encoder reranking, hallucination control, barge-in handling, and per-stage latency instrumentation.

    Python