Who is Kartik Malhotra?
Kartik is the founder and principal engineer of AIVETech and builder of Soil. His work includes SaaS application logic, integrations, source-access boundaries and reviewed workflow actions. See the implementation evidence and current limits.
How you work together
Kartik leads the scope, architecture and implementation. An engagement defines one feature, the integrations it needs and how its results will be tested. Progress reviews, documentation, deployment approval and support responsibilities are agreed in the proposal.
Before AIVETech
Kartik led AI at TheraNow, a digital health company, and worked as an engineer at KredMe, a consumer fintech product. He did not found either company; the work was done as an employee. Both are named here because they are where the engineering judgement came from, not as AIVETech client references.
That history shapes what AIVETech scopes. Health and fintech products are where a confidently wrong answer has a cost, where data boundaries are examined rather than assumed, and where a feature that demos well but fails on the tenth edge case does not ship. It is also why the service pages on this site describe evaluation, permission scopes and failure handling at more length than they describe models.
What he builds, and how
The work is agentic systems that take bounded actions through approved tools, voice systems that handle calls end to end, retrieval over a company's own documents, and automation across existing business systems. In every case most of the engineering sits around the model call rather than in it: reaching the data, defining which actions are permitted, handling the cases where a provider or a downstream system fails, and keeping cost per request predictable.
The approach is to narrow the problem until it can be evaluated. A bounded job with a written definition of completion can be tested, scored on every change and argued about with evidence. An open-ended assistant cannot. That is the single decision that most often separates an AI feature that reaches production from one that stalls in a demo.
Soil
Soil is Kartik's own product, built and run in-house: it records a customer promise, assigns an owner and deadline, reviews the work behind it and separately confirms delivery reached the customer. Its core commitment tracking is rule-based rather than model-driven, which is itself the point, since the reliable part of the problem did not need a model. The implemented capabilities and current limits are documented, and the public demo uses fictional data.
Writing
Kartik publishes engineering notes on the Aivetech blog, including the latency budget of a production voice AI agent, why naive RAG fails in production, and designing guardrails for autonomous AI agents.
Contact
Reach Kartik at hello@aivetech.in, on LinkedIn, on Instagram, on GitHub, or share a project brief. Note: this page is about Kartik Malhotra, the AI engineer and founder of AIVETech — not to be confused with other professionals who share the name.
Work with Kartik on your AI system.
Share the problem, your product and the result you need. Kartik reviews the brief before proposing next steps.
Discuss an AI Project →