Gaurav Chauhan
AI Engineer | Generative AI & Machine Learning
I design and build intelligent systems using large language models, generative AI, and end-to-end ML pipelines. I turn business ideas into working AI products.
Featured AI & GenAI Projects
Real-world projects in LLMs, Generative AI, and Machine Learning.
Problem
A full fledged ai astrologer agent as telegram userbot, With agentic memory.
My Role
Designed bot api, prompt engineering, chat agent, agentic memory and token embedding.
Key Features
- •Birth chart generation from date, time, and location
- •Daily horoscope predictions for all zodiac signs
- •Personality analysis based on astrological principles
- •Conversational memory for personalized interactions
- •Real-time planetary position calculations
- •Zodiac compatibility matching
- •Multi-language support (Hindi, English)
Impact
Enabled 24/7 accessible astrology consultations, reducing wait times from days to instant responses.
Problem
Live surveillance-based fire and smoke detection system with alert API
My Role
Built a custom trained yolo model and created a live alert api using the model and cctv cameras
Key Features
- •Real-time fire and smoke detection from live CCTV streams
- •Custom YOLO model with high accuracy (>90%) on fire detection
- •Multi-camera support for comprehensive monitoring
- •Instant alert API for integration with alarm systems
- •Bounding box visualization with confidence scores
- •24/7 automated surveillance without human monitoring
Impact
Achieved 92% detection accuracy with <1s response time, enabling early fire detection before spread.
Problem
Generated surreal photorealistic UGC using Flux.1 D, Qwen Image Edit, WAN2.2, Infinite Talk, and Vibe Voice through ComfyUI.
My Role
Built extreme cost-effective surreal generation pipeline using Flux.1 D, Qwen, WAN2.2, Infinite Talk, and Vibe Voice with custom LoRAs.
Key Features
- •Extremely surreal generation using Flux.1 D and WAN2.2 models
- •Advanced image editing with Qwen Image Edit for surreal transformations
- •Infinite Talk for realistic lip-sync and audio-video synchronization
- •Vibe Voice for natural voice generation in video content
- •Custom LoRA fine-tuning for hyper-surreal brand aesthetics
- •Cost reduction and minute control through open-source model integration
Impact
Achieved extreme cost reduction compared to manual content creation, influencer collabs and hiring creators.
Problem
Using old pc, hdd and open source softwares like trunas, nextcloud, immich, build a home server for all my personal files and photos
My Role
Designed and deployed home server stack with TrueNAS for OS, Nextcloud for file sync, and Immich for photo management.
Key Features
- •TrueNAS Scale with ZFS for enterprise-grade storage management
- •Nextcloud for self-hosted file sync and collaboration
- •Immich for private photo backup with AI face recognition
- •Automated backups and device synchronization
- •Zero monthly subscription costs vs. cloud services
- •Remote access to personal files and media library
- •RAID configuration for data redundancy and protection
Impact
Eliminated $120+ annual cloud storage costs while maintaining full data ownership and privacy with enterprise-grade reliability.
Skills & Expertise
A comprehensive toolkit for building AI systems at scale.
Experience & Education
Professional Experience
Training and deployment of ML models and gen ai solutions for Agriculture at enterprise grade
- ▸Analyzed soil and weather datasets to develop actionable insights for agricultural planning.
- ▸Built predictive models in Python for fertilizer and seed recommendations based on environmental data.
- ▸Designed and deployed computer vision models to detect crop diseases from field images.
- ▸Developed and integrated a LLM-powered chatbot to answer farming queries dynamically.
- ▸Created backend services and REST APIs for real-time model interaction and data flow.
Education
Focus: Machine Learning, Artificial Intelligence
Focus: Data science, ML, Deep Learning, NLP, Computer Vision