A data analytics, machine learning and AI company, working out of Belagavi with clients worldwide.
TECHWORKSLAB Private Limited is a data analytics, machine learning and AI company. We work with organisations that have accumulated a great deal of operational data and comparatively little benefit from it.
Two domains sit under one company. This site covers the AI products and solutions work. Clinical, regulatory and submissions services run under techworkslab.com. Same people, different problem.
We would rather tell you in week one that machine learning is the wrong tool for a problem than bill for six months of discovering it. Most of the value in this field comes from framing the problem correctly and getting the data into a usable state, neither of which is glamorous.
Founder and director of TECHWORKSLAB Private Limited. Lead R developer and technical consultant, and a research scholar at IIT Ropar.
Author and maintainer of edgemodelr, a package on CRAN that runs large language models locally in R from GGUF files through llama.cpp, with no cloud API and nothing leaving the machine. That is the same constraint the benchmarking work meets constantly: a client who cannot send their cases to somebody else's endpoint.
Also author of the R4SUB framework, five packages on CRAN that score how ready a clinical submission is: a core evidence contract, an FMEA-based risk engine that computes risk priority numbers, a traceability engine, and regulator-specific profiles carrying their own thresholds for the FDA, EMA, PMDA, Health Canada, TGA and MHRA.
That is the same instrument as the benchmarking product, pointed at a different problem. Agree the thresholds with the people who own the decision, score the evidence against them per category rather than as one average, and end with a readiness answer somebody can act on. R4SUB does it for a submission and the benchmark does it for an agent.
Writes the engineering notes here, and wrote the four curricula and six workbooks published under EDUSHARK TRAINING: 60 weeks of lessons and 1,418 pages, in which every code example was run before publication and the printed output is what the machine returned, including the runs that undercut the technique being taught.
Rather than a list of adjectives, the three engineering notes on this site and the case studies from the clinical practice are the evidence. They describe real systems and state what changed.
Three sites, one registered company. Worth stating plainly, because a contract has to name one of them.
We will say plainly whether machine learning is the right tool for it.
Start a conversation