Competency Name Generative AI

Broad Definition: This competency defines the ability of an associate to understand and potentially use Generative AI in areas like problem-solving,
decision-making, natural language processing, autonomous systems, and personal assistance.
Criteria E0 E1 E2 E3 E4
I Criteria for defining proficiency vis- N/A N/A N/A N/A N/A
à-vis knowledge & skills

1. Precondition None Have working experience across one or more areas of Gen Have advanced knowledge in using AI models with Have strong working knowledge in
[Expressed as a prerequisite - in terms
Have a fair understanding of Deep AI. a strong working experience on NLP, Deep Advanced AI Models and in one or more
of academic qualification AND / OR
industry experience (whichever is Learning/NLP concepts. Learning and Generative AI areas. domains. Has provided technical consulting
applicable) AND / OR relevant on Gen AI areas.
knowledge in any other competency]

2. Knowledge Having a basic understanding of Having Hands-on knowledge of: - Advanced knowledge in using AI models - Provide technical consulting - Expertise in domain-specific LLM
[Expressed as Area of knowledge and
- Gen AI terminologies and its capabilities, - Using advanced models - Usage of public LLMs and handling embeddings in pre- - Build Solution architecture and orchestration of creation
not skill, includes knowledge of
process, methodologies, concepts and - History of various AI models, - Multi-modal models processing. LLMs - Building RLHF capabilities
tools] - Gen AI applications in different industries, - Performing model evaluation across the bias, - Fine-tuning of the models addressing improvements such - Build Multi-modal solutioning like a combination of - Handling complex business cases
- Limitations and risks, toxicity, accuracy, etc., as accuracy, low latency, etc audio, video, and text - together, - Experimenting with alternate use of
- Do's & Dont's while using these models, - Understanding the methods and metrics - Handling AI responsibly -Perform Performance Engineering model architectures such as Adversarial
- Prompt engineering associated with the models - Cost Prediction of models networks, transformers, diffusion models
- Knowledge of hyper-scaler models

3. Experience None 3 months 1 year and more 2 years and more 3 years and more
[Expressed as duration in months /
years wrt exposure, including breadth &
the depth of assignments; also beyond
project experience]
3.1 Type of relevant experience None Should have worked on prompt engineering Should have developed applications consuming GenAI APIs
[Type of assignments/ areas /outcomes
the professional has worked on; total
time to be spent on each function] Should have used one or more hyper-scaler Should have worked across various LLMs and used them as
models per the business needs N/A NA

3.2 Beyond project development NA Should have contributed to E0 level training Should have conducted at least Two E2-level Should have evaluated associates for
experience [Includes
programs/Content curation Should have contributed to training with at least two end-to- training programs a year. awarding E3 level
[1) Relevant interdepartmental activities
2) Involvement in corporate initiatives]
end business case studies toward training material at E1
level
Should have contributed towards at-least Should have groomed at least one
Should have contributed with 2 E1 level training sessions a Three solution proposals a year associate in Solutioning
year

4. Asset Creation NA Should have worked on creating Prompt Built small LLMs and use cases that can be used for Should have worked on one or more whitepapers Worked on at least one white
[Expressed as any reusable and
Libraries conducting training programs on Business implementations/solutions utilizing paper/patent/reusable framework.
deployable assets created in the
knowledge management system of the GenAI
organization]
4.1 Process frameworks defined NA NA NA NA NA
[Articulated as the process
improvement designed OR a new
process framework designed]

4.2 Tool creation NA NA NA NA NA

[Expressed as a system or function tool
designed OR created towards
improving business delivery / learning]
4.3 Reusable & deployable assets NA Should have worked on creating Prompt Built small LLMs and use cases that can be used for Should have worked on one or more whitepapers Worked on at least one white
created
Libraries. conducting training programs. on Business implementations/solutions utilizing paper/patent/reusable framework.
[Expressed as reusable and deployable
assets created (like those in Mighty &
GenAI.
Knowmax]

AND/ OR
No. of training material created NA NA Should have contributed to training with at least two end-to- Should have enabled Training Content and Use Should have evaluated associates for
[Expressed as the number of relevant
end business case studies towards training material at E1 Cases for at least two programs a year. awarding E3 level competency.
training material created towards
development of a particular body of level.
knowledge or skill]

5. Industry contribution NA NA NA Should have worked on one or more whitepapers Worked on at least one white
[Expressed as the contributions made
on Business implementations/solutions utilizing paper/patent/reusable framework.
to the industry. This criteria should be
measured by mechanisms for e.g.: TCS GenAI.
GEMS]

TCS Confidential
5.1 Recognized by Peers NA NA NA NA NA
[Expressed as the recognition given by
the peer community and should be
measured by appropriate mechanism
such as TCS GEMS]

AND/ OR NA NA NA NA NA
Recognized by a Parent Body NA NA NA
[Expressed as the contribution
recognized by a renowned and
recognized parent body]

AND/ OR NA NA NA NA NA
Invited as Speakers NA NA NA Should have represented the Organization in Should have represented the Organization
[Expressed as recognition from an
industry forums and AI conferences in industry forums and AI conferences
external body when it extends an
invitation to the speaker to address a
recognized forum]
II Assessment / Evaluation NA NA NA NA NA
Methodology vis-à-vis Proficiency
Level
1. Certification As per Learning Object (LO) As per Learning Object (LO) NA NA NA
[Expressed as achievement of external
programs - certifying a part of the
proficiency level, Achievement of
internal programs - internal trainings
attended]
2. Evaluation Test As per Learning Object (LO) As per Learning Object (LO) NA NA NA
[Expressed as the assessment
mechanism used to assess the
individual for having achieved the said
proficiency level. (for eg.,Type of tests
for each proficiency should be defined-
like for E0 level it can be only self
assessment, for E1 it can be panel
review etc)]
Version History
Project Details : HR-TD-Technology
Unit : HR1.0
Parent Project - HR-Parent1
Project Owner/GL : Anupama Muralidhar (119131)
Amendment History:
Version Defined By Approved By - Role & Group Facilitated By Date Remarks

Version 1
Kumaresan. S (243823)
SAI KUMAR SOMANCHI( 567290) Anupama Muralidhar (119131) 21-Jul-23 -
iEvolve Data Management Team

Rajan Pillay (471334)

Vijayalakshmi s( 1051644)

TCS Confidential