Skip to content

Ritvij Kumar Sharma

Principal Engineer, Kendal AI

  • Agents
  • Engineering
  • AI Infrastructure
Ritvij Kumar Sharma, red-ink portrait in the thehype style

The take

Developers have become a bottleneck when it comes to reviewing code generated at such fast pace.

If senior engineers don't read the code being put in production, applications will break – race conditions, scalability issues, something the AI just missed. The answer boils down to one thing: trust. How do we trust our agents more? That leads us to building agentic coding systems and harnesses in such a way that, even without reading code, we can trust the work – system-design discussions with your agents up front, and verification systems around the harness so agents can verify their own work.

— Ritvij Kumar Sharma

Editorial bio

Ritvij Kumar Sharma is five years into a career most engineers spend decades building, and his central provocation is that reading code is already optional. As Principal Engineer at Kendal AI – where he built the CRM used by top real-estate brokerages across MENA from 0 to 1 – he spends most of his day inside agent harnesses and agent memory systems, directing models rather than editing diffs.

What saves him from the bottleneck he's first to admit is real: trust engineered up front. System-design conversations with the agent before any code lands, verification systems around the harness so agents can check their own work. He published on it (IEEE ICSA 2023), ships on it (production real-estate tooling across the region), and treats "trust the agent" as an engineering problem, not a sentiment.

Worth following if you want to see what senior engineering looks like when the code stops being the artifact.

Key facts

based in
UAE (Indian national)
current
Principal Engineer, Kendal AI
prior
Full-stack engineer across mobile, web, blockchain and LLM apps · ~5 YOE
focus
Agent harnesses · agent memory · agentic coding workflows · system design
research
IEEE ICSA 2023 (DOI 10.1109/ICSA-C57050.2023.00055)

Featured work

  1. engineering

    Kendal AI (Principal Engineer)

    Built the Kendal AI CRM from 0 to 1; now used by leading real-estate brokerages across MENA.

  2. research

    IEEE ICSA 2023 paper

    DOI 10.1109/ICSA-C57050.2023.00055.

  3. engineering

    Agent harnesses & memory systems

    Ongoing public exploration of what makes agentic engineering competent vs. chaotic.

Share this profile

On air · radio.thehype.news ↗ AI news radio — 24/7