IIT Delhi grads build, sell AI startup in 4 months; now raise $1 million for new venture

Anirudha Kulkarni and Pratyush Saini have raised $1 million for ByteAsk, their new AI coding startup for C and C++ developers. The venture applies lessons from their quick LawSutra AI exit to specialised, reliability-critical software work.

In Short

• ByteAsk will serve sectors needing accuracy, speed and dependable software systems
• The startup works across India and the US, targeting specialised engineering teams
• Founders say generic copilots often struggle with legacy-heavy, hardware-linked development

Two IIT Delhi computer science graduates who built and sold an AI legal research startup within four months have now raised $1 million in pre-seed funding for their second venture, which is building AI coding agents specifically for C and C++ developers.

Anirudha Kulkarni and Pratyush Saini, who previously co-founded LawSutra AI, have launched ByteAsk, a San Francisco-based startup focused on AI-assisted software development for industries where reliability, performance and accuracy are critical. ByteAsk has raised the funding from Y Combinator and Entrepreneur First, along with angel investors from global quantitative trading firms.

The startup was founded in June 2026 and has also been selected for Y Combinator's Fall 2026 batch. It operates across the US and India and is targeting developers working in areas including automotive, aerospace, robotics, high-frequency trading, embedded systems and semiconductors.

WHY C AND C++?

ByteAsk is being built around what the founders describe as a gap in existing AI coding tools. While generic coding assistants have increasingly become useful for languages such as Python and Java, the founders argue that C and C++ development presents different challenges because of large legacy codebases, debugging, testing, performance requirements and hardware dependencies.

C and C++ are widely used in systems where software performance and reliability are important, including vehicles, trading systems, chips, operating systems and other embedded applications. ByteAsk says its coding agent is designed to build, debug and test changes using the actual development toolchain before presenting them to developers.

The platform combines an agentic coding environment with technical documentation, standards and specifications relevant to C and C++ development. The founders say this approach is intended to provide the AI with additional context and help reduce errors in specialised coding tasks.

“Building LawSutra taught us that models are only as useful as the context, verification and workflows around them. At ByteAsk, we are applying that same philosophy to code,” Kulkarni said.

EXPERIENCE FROM FIRST STARTUP

Before ByteAsk, Kulkarni and Saini founded LawSutra AI, an AI-powered legal research startup. The company used semantic search to help lawyers find relevant legal information based on the meaning and context of a query rather than relying only on keywords.

LawSutra attracted more than 5,000 lawyers before being acquired by legal research company Manupatra four months after its founding, according to reports.

The experience shaped the founders' approach to their second company. They are again focusing on a specialised domain where they believe generic AI systems can struggle without additional context and verification.

Both founders studied computer science at IIT Delhi and have experience working on AI and systems infrastructure.

Kulkarni holds BTech and MTech degrees in computer science from IIT Delhi. Before ByteAsk, he worked on low-latency C++ systems at Optiver and on markets engineering at Quantbox Research. He also built a Python-to-C++ optimisation system at Quantbox, according to ByteAsk's website.

Saini holds a BTech in Computer Science and Engineering from IIT Delhi. He previously worked on AI infrastructure at ThirdAI and was a founding engineer at AI security company SimbianAI.

WHERE WILL NEW FUNDING BE USED

The new funding will be used for research and engineering, GPU computing, training data and enterprise infrastructure. ByteAsk also plans to develop security, privacy and on-premises deployment capabilities for companies that may not want sensitive code or data processed on external systems.

The startup says its initial customer focus will be on sectors such as high-frequency trading, automotive and embedded systems, before expanding into other areas.

ByteAsk is also working on its own post-training approach for C++ and plans to release a specialised model in the coming months, according to the company's founders.

The two IIT Delhi alumni are now attempting to apply lessons from their first startup to a different problem: building AI tools that can work reliably with some of the world's most complex and performance-sensitive software systems.

- Ends