Headline

Agentic AI Engineer @Threetone Labs | Partner @hack47 | Building @Collision | Ex-xAI | Hackathons @bits&bytes Foundation

About

I build advanced agentic AI systems for long-running enterprise use cases, orchestration, persistent memory, self-improvement loops, and task-specific learning. My current focus is SkillLoop, a standalone self-improvement runtime for harness for AI agents. It sits beside any agentic harness, ingests completed traces, evaluates them, proposes memory and skill updates, requires human review before applying anything, and exports fine-tuning-ready SFT/DPO datasets. Deliberately decoupled from the harness itself for versioning, auditing, and enterprise-grade provenance. Before this: Worked at xAI training Grok for multilingual training and voice AI at ThreeTone Labs, product engineering at SH1P, and also worked on a multi-agent e-commerce optimization systems and anomaly detection system in the milk supply chain for a large consumer brand in India. Also built polymarket weather arbitrage agents. Also executed the world's largest hackathon (26K registrations, 4K+ on-site) and scaled a YouTube channel to 100K+ subscribers.

Location

Jaipur

Work experience

Agentic AI engineer at ThreeTone Labs, Feb 2026 - Present · 7 mos, Remote

Product Engineer at SH1P, Feb 2026 - Present · 7 mos, India, Worked on product launches and iteration cycles, collaborating across growth and product to ship features quickly. Built internal tooling and contributed to product development with a focus on zero to one execution and user feedback.

Venture Scout at LvlUp Ventures, Jan 2026 - Present · 8 mos, Remote, Source and evaluate early-stage startups for potential investment. Conduct market research and due diligence, identify high-potential founders, and connect them with investors and resources. Provide guidance on pitch preparation, fundraising strategy, and network introductions.

Multilingual Tutor at xAI, Jan 2026 - Apr 2026 · 4 mos, Remote, Training Grok for multilingual Audio. And creating cutting-edge datasets for training, benchmarking, and advancing large language models.

Founding Growth Lead at Maximally, Aug 2025 - Mar 2026 · 8 mos, India

Founder's Office at Maximally, Apr 2025 - Aug 2025 · 5 mos, At Maximally, I serves in the founder's office—where startup energy meets soulful execution. Helped in: operations, outreach, storytelling, and scalable systems that support the dreams of young entrepreneurs. From pitch decks to podcasts, I help shape ideas into impact.

Education

Bachelor of Science - BS, electronic systems, Indian Institute of Technology, Madras

Public links

github: https://github.com/lamenting-hawthorn