Artificial intelligence dey change how people dey find new medicine fast fast as drug company dey find way to reduce time and money and still make sure say new molecule fit turn to real treatment. More than 200 startup don enter this AI for drug research space. Converge Bio wey dey between Boston and Tel Aviv don show now as one of the startup wey get better funding for this area.

The company don raise 25 million dollar Series A money wey pass the target. Bessemer Venture Partners lead the round and TLV Partners plus Vintage Investment Partners join. Some senior people from big tech company like Meta OpenAI and Wiz also put money to show say tech industry dey watch how AI dey enter biology.

Converge Bio idea simple but big. Dem wan build generative AI wey dem go train directly on biological sequence like DNA RNA and protein then connect am into the normal work wey pharma and biotech company dey do. Instead make researcher dey do trial and error for wet lab all the time dem fit use Converge model design test and adjust molecule for computer first then carry only the best ones go lab.

Converge Bio CEO and co founder Dov Gertz talk say drug development get clear stages from target identification and discovery to manufacturing clinical trial and beyond and for each stage dem get experiment wey Converge fit support. E talk say their platform dey expand across all these stages to help new drug reach market faster.

Instead make dem sell only one big model Converge Bio dey sell wetin dem call systems. Na toolset wey join many AI model together inside one workflow wey researcher fit use immediately. For now dem don launch three system one for antibody design one for protein yield optimization and one for biomarker and target discovery.

The antibody design system show how their style be. No be just one generative model dey throw out antibody candidate. Na full pipeline. First generative model go suggest new antibody sequence. After that predictive model go check those sequence for important property like stability how easy e go be to manufacture and how e fit work for body. Last last physics based docking engine go simulate how each antibody go take interact for three dimension with the target so dem fit know the best ones to push forward.

Gertz talk say the real value dey inside the whole system not just one model. Customer no need dey join different model by themselves. Dem just get ready made system wey dem fit plug into their normal work.

This new money come after earlier seed round and e close one period of fast growth for the two year old company. Converge Bio don sign about 40 partnership with pharma and biotech company and dem dey run about 40 active program for their platform. Their customer dey US Canada Europe and Israel and dem don start to enter Asia market too.

Number of staff don grow from small single digit to several dozen people and dem share team between Boston and Tel Aviv office. As the company dey grow dem don begin release case study to show wetin their technology fit do for real work.

For one collaboration Converge talk say dem help one partner increase protein yield by about four to four and half times for just one computational round. For that kind field na small small improvement dey common so this one big. For another project their platform generate antibody wey get very strong binding affinity for single nanomolar range and that kind performance fit be key to turn candidate molecule into real drug.

All this dey happen as interest for AI drug discovery dey rise. Big pharma company like Eli Lilly don partner with Nvidia to build powerful supercomputer for drug research. The AlphaFold team from DeepMind wey dey predict protein structure don also get big recognition for science and this one don confirm say AI don become serious tool for structural biology.

For Gertz mind all these things show say life science work don dey change seriously. E talk say this time fit be the biggest money opportunity for the sector because people dey move away from pure intuition and trial and error go data driven molecular design.

When Converge Bio start many potential partner no too believe say AI fit really help for this kind complex area. But as more case study from startup and academic lab don show say AI fit design better molecule optimize process and find new biological insight that doubt don reduce well well.

Still the field get real challenge. Large language model wey people dey use for chatbot don also enter drug discovery because dem fit analyze biological sequence and scientific paper. But these model dey hallucinate sometimes give confident but wrong answer and for biology that mistake fit cost plenty time and money.

Gertz talk say for text you fit see hallucination quick but for molecule to check new compound fit take weeks so the cost high. To reduce this risk Converge Bio dey pair their generative model with predictive model wey act like filter. These predictive model go score and rank new molecule before anybody touch lab bench. The aim no be perfection but to cut risk reduce dead end and avoid expensive experiment on molecule wey no get chance from beginning.

Gertz also dey separate Converge core technology from the text based LLM wey people dey always talk about. E agree with some AI researcher wey talk say language model alone no fit really understand biology. E talk say dem no dey rely on text model for core scientific understanding. To really understand biology model must train on DNA RNA protein and small molecule.