Looking for a job

Siddheswar User

@siddheswaruser

Message

I build production-ready generative AI, RAG, multimodal vision, and real-time voice agent systems.

India

Message

What I'm looking for

I'm looking to build and productionize generative AI systems across LLM agents, RAG, multimodal AI, and real-time voice or vision infrastructure.

I've built production-focused generative AI systems at TruGen AI, including a real-time multimodal vision pipeline using SmolVLM for TruGen Avatar to detect objects and answer questions about its environment.

At FutureSmart AI, I optimized Knowledge Base and NL2SQL microservices, reducing end-to-end latency by 35% while maintaining 90%+ uptime. I also reduced AI pipeline errors by 20% through automated validation and debugging workflows for client-facing chatbots.

At Enercent Technologies, I architected a Gemini API and ChromaDB-powered PDF intelligence platform that reduced manual query response time by 60%. I build across LLM agents, RAG, multimodal AI, real-time audio and video infrastructure, and developer tooling.

Experience

Work history, roles, and key accomplishments

Current

Gen AI

Current

TruGen AI

Jan 2026 - Present (8 months)

Engineered a real-time multimodal vision pipeline using SmolVLM, enabling TruGen Avatar to detect objects and answer natural-language questions about its visual environment. Built and documented reusable AI components through a developer playbook, accelerating internal adoption and reducing repetitive engineering questions by an estimated 30%.

SmolVLM Multimodal AI Pipecat Developer Tooling

Gen AI

FutureSmart AI

Sep 2025 - Dec 2025 (3 months)

Optimized 4 modular AI microservices, including Knowledge Base and NL2SQL systems, by refactoring inter-service communication and reducing end-to-end latency by 35% while maintaining 90%+ uptime. Reduced AI pipeline error rates by 20% through automated validation and debugging workflows for production client-facing chatbots.

Microservices NL2SQL Knowledge Base Validation Debugging

Gen AI

Enercent Technologies

Jan 2025 - Aug 2025 (7 months)

Architected an enterprise PDF intelligence and RAG platform using Gemini API and ChromaDB, enabling context-aware question answering over business documents and reducing manual query response time by 60%. Built an automated data-harvesting pipeline using BeautifulSoup and Selenium, collecting and normalizing information from 10+ sources and improving retrieval efficiency by 80% over manual collect

Gemini API ChromaDB RAG BeautifulSoup Selenium YOLOv8 OpenAI OCR