Niwant Salunke
AI Engineer @ Theoremlabs.io
About
I build AI systems that don’t just generate text — they take action.I’m currently building real-time AI meeting agents that join live enterprise calls, transcribe, reason, and trigger downstream workflows across APIs and internal systems.My focus: agentic workflows, event-driven architectures, and AI systems that operate inside real-world production environments.Previously:• Built AI meeting agents using Recall.ai + ElevenLabs• Designed Azure-based event-driven orchestration for Siloed Systems• Fine-tuned VLMs (CLIP / BLIP) for multimodal tasks• Built MCP tooling to auto-generate admin dashboards from schemas• Increased test coverage + deployment velocity at Software AGI like ambiguous problems, small teams, and shipping fast.If you're building AI-native products, I’d love to contribute.
United States
Charlotte Metro
Computer Software
Swift (Programming Language), Virtual Reality (VR), iOS, 3D Modeling, Knowledge Graphs, Retrieval-Augmented Generation (RAG), Text-to-Speech, Real Time System Design, Agentic Workflows, LORA, Fine Tuning, VLM, QLORA, Model Context Protocol (MCP), Large Language Models (LLM), Engineering, Front-End Design, Three.js, Architectural Design, 3D Animation
Experience

AI Engineer
Charlotte, NC
- Built a real-time AI meeting agent that autonomously joins enterprise meetings, transcribes live conversations, generates structured outputs, and responds context-aware in-session. - Architected event-driven workflows using Azure Logic Apps + Event Grid to handle transcription routing, agent reasoning, and post-meeting automation. - Designed secure API gateways with Azure APIM to support asynchronous enterprise integrations. - Contributed to banking automation PoCs enabling legacy RPG system interoperability via Azure SQL MI and workflow orchestration.

Research Assistant — AI4Health (BAMM)
Charlotte, NC
- Built motion processing pipelines converting .bvh skeletal data into normalized joint-space representations for consistent cross-model alignment and downstream analytics. - Developed motion interpolation algorithms to synthesize temporally consistent in-between frames, improving realism and continuity in generated 3D animation sequences. - Optimized PyTorch inference pipelines for real-time deployment, enabling web-based model interaction with low-latency motion synthesis. - Contributed to deployment-ready tooling bridging research-grade motion models with production-facing web interfaces.

Research Assistant — UQ Labs (VLM / MCP / Multimodal)
Charlotte, NC
- Fine-tuned CLIP and BLIP-based visual-language models for spectrogram classification using multimodal embedding alignment. - Designed zero-shot and few-shot evaluation pipelines leveraging vector similarity and cross-modal retrieval techniques. - Built scalable inference workflows for LLM–VLM integration, optimizing deployment for multimodal production systems. - Developed Model Context Protocol (MCP) tooling that auto-generates React-Admin dashboards directly from backend schema definitions, enabling rapid internal tooling generation.

Software Engineer 1
Vasai Virar, Maharashtra, India
- Engineered and deployed scalable features for WebMethods.io integrations using React, Node.js, and MongoDB, achieving a 90% improvement in unit test coverage from 32% to 61%. - Architected and optimized CI/CD pipelines with Jenkins and GitHub Actions, reducing deployment cycles by 75% while ensuring zero-downtime deployments. - Refactored and modularized complex React components, improving reusability, maintainability, and reducing front-end page load times by 50%. - Identified and resolved critical security vulnerabilities by implementing Content Security Policy (CSP) headers, sanitizing user inputs, and addressing cross-site scripting (XSS) and SQL injection risks, enhancing overall system security by 70%. - Created custom scripts for data migrations with backup and restore capabilities, ensuring seamless transitions during platform upgrades. - Spearheaded the integration of RESTful APIs and GraphQL services, improving data query efficiency by 20% and enhancing application interoperability. - Mentored onboarded junior developer, standardizing practices for version control, code linting, and CI/CD workflows to boost team productivity by 30%.
Niwant Salunke's Contact Information
Phone
Find the Right Leads
Find Verified Contact Data
What LeadContact does well
Find verified emails, phone numbers, and decision-makers with 98% accuracy.
Find Leads
Find the right people by company, role, industry, location, and more.
925M+ professional profiles

Find Emails
Access verified email addresses for your target contacts.
657M+ emails

Find Phone Numbers
Get cross-validated phone data from multiple top sources.
239M+ phone numbers

More Accurate. Lower Cost.
Find contact data in 1 tool with 98% accuracy
LeadContact integrates leading enrichment tools to deliver more accurate contact data—without paying for each one.
Great conversations start with the right contact.
It’s time to find yours.



