My Take on the AI Stack

   |   5 minute read   |   Using 1015 words

When machine learning started taking off around 6 years ago, before the advent of ChatGPT, everyone started chasing after the title of “machine learning engineer”. If you looked at their resumes and the job postings that targeted them, you would have found a variety of skills across the stack, ranging from data science to MLOps to GPU programming.

Now, an even more loaded term is floating around the industry: the “AI engineer”. My frustration from back then remains the same. Whenever I’m asked about what I’m working on or what paper I’ve read related to the AI field, I have a hard time setting the context of where exactly the topic falls.

The following is a top-down view of my take on the LLM-focused AI stack.

TTTTTTTTTTTTiiiiiiiiiiiieeeeeeeeeeeerrrrrrrrrrrr1198765432110LAALLLDMMGIHaggLLLeaLPnayeeMMMecUtrennphFedrttITAirPrwsnrrLnarmaHfaceemoera/eihaegderrnirLwri/nAeitneoaaNegnneiarmtesecgcnrkmetsne/tgnsiwet/FuinRosiSirngercenegpk/reriAvtenMpiusgupnnellgintinticga-atatigioeonnnsts(sIyRs)temsCLGMRFCPMSVTPNtCPGCoLeCelorumirCeoUTPanMnPtanelasaAarDXUccOerstttlinrcAheprSihirilosehipsaeenamnfCs.PTSnttrvAuioL-oltcriRgsovatondaLru-oolAPreltuianamsNmgeMr-r-esnlgneteprNoEsAngugreiiaIVmvutBLausrglmRDLpagiaLgaihemPitlPmotFMegnbiUnurenciseCgo(nSkOaonhnM-NrPgAptmtieoANySTtopeKntMdcsASTMSPRirtdVguietueooUDmnxlitardMiLEGCitsoRorceAzoneaPnunNechlSaogncagrNnh)ytpiehgeMmscsSisnree-uooTtyoeadDolddXrosnetAifteLeLiltriSts-ilSrAtie(iopttEhsTsocmPnneeixeM(n-rgcnlpa(sJAooutledVArnmliarLXRc-pSaottLAE)ohMtetnisasmCiomiot)bMtdOavneeDeupneQndeclstutdete)iDaipucenAnSr(cttgpeSSoitesOedzemeMaianod)rnttdcgiiehoolnns


© Sridhar Nandigam