Section
Machine learning, generative AI, and what it means for work, culture, and power.
Sarvam AI opened an Indic voice API for Bengaluru call-centre pilots, targeting Kannada-Telugu-Hindi code-switching on outbound collections and utility bots with on-device latency budgets.
HCLTech deployed an on-premises generative AI stack for public-sector data centres in Noida, giving ministries retrieval-augmented assistants that keep citizen records inside sovereign racks rather than public cloud APIs.
Nasscom published Hindi–English LLM benchmarks timed to MeitY’s GenAI sandbox, scoring retrieval, summarisation, and code-switch handling so vendors cannot hide weak Indic performance behind English-only demos.
MeitY opened a GenAI sandbox that lets startups train and evaluate models on curated Indian-language datasets surfaced through BHASHINI, with audit logs aimed at government procurement.
Wipro embedded a document-retrieval copilot into renewed BFSI managed-services statements of work, routing loan and KYC packets through private India regions with RBI audit logs on every prompt.
Sarvam AI released Hindi-tuned legal summarizer weights for district court filing desks, pairing on-prem inference with eCourts document schemas as MeitY pushes sovereign AI pilots in judiciary IT stacks.
