Princeton Strategy Group
Princeton Strategy Group
Digital Transformation and AI/ML Adoption,Climate Risk Advisory, Risk, Finance and AML/Fraud Modeling.

We can help you integrate cutting-edge AI/ML, LLMs and NLP in your key business process workflows to better manage your obsolescense risk and stay ahead of the competition.

We have experience implementing newest generation foundational models fine-tuned and optimized for financial domain. Our experts can develop quick protypes using both open source and behind-the-paywall models to deliver the high performance models to suit your business case without compromising your data security.

Key Challenges


  • Proprietary development of a foundational model is prohibitively expensive due to the enormous compute power requirements.
  • Fine-tuning an LLM (Llama 2) to suit the business domain is much more feasible but organizations lack the necessary skill set.


  • Most organizations do not have the data organized in place for real-time LLM inference. Even structured data needs to be reorganized using vector databases for real performance.


  • While the usage of APIs for Large Language Models (LLMs) such as ChatGPT has skyrocketed in the past year, security and lack of transparency have been major concerns. For example, the inadvertent leakage trade secrets or sensitive personal information used as training data which may not be transparent.


  • Most of the LLMs today are too large for even a big organization’s typical use. More medium sized language models are required to mitigate the inefficiencies.

Our experts can help you assess your current data architecture readiness for AI adoption and recommend the leading tools and technologies to remain competitive.

Prompt Engineering

  • Persistent Prompt Engineering can fine-tune the models to custom domain. LangChain is an excellent tool to explore this.

Retrieval Augment Generation

  • Retrieval Augmented Generation (RAG) allows knowledge update of LLMs via the model’s knowledge base via information fed into the LLM via the input prompt.


  • Middle-ground approach but it will be custom tailored to own data.

Proprietary Model

  • Using Transformers and other deep learning techniques to train smaller models to suit the business cases on specialized hardware.

Practical Approaches

We have been helping several financial institutions in quick exploration of some of the practical approaches. For one of our clients, We prototyped a LangChain based interface to prompt engineer LLM responses based on firm's analyst and economist reports, to generate accurate summaries and Q&A to proactively enagage the cutomers.

Data Readiness

Data Architecture

  • Legacy architectures impede speedy AI adoption and need to be replaced by newer generation technologies. A total rearchiteture might pay huge dividends in the long run.


  • Hybrid Transaction/Analytics Processing databases can serve AI models more effectively.

Structured and Unstructured

  • NewSQL can handle data through SQL servers s(plain old SQL) as well as NoSQL through HTTP/REST.
  • Use Vector databases for unstructured data. Many open-source technologies are available for easy adoption.

Avoid costly ETL

  • Costly ETL processes should be minimized or avoided to sustainably serve AI models in real-time.

Please feel free to contact us to learn how our experts can help you assess your current data architecture readiness for AI adoption and recommend the leading tools and technologies to remain competitive.


Sustainability- Humanity will meet current economic needs without overburdening the natural environment to ensure future generations will be able to meet their economic needs.

SDG - Stands for Sustainability Development Goals. 17 goals defined and subdivided into 169 targets. Environmental (E) goals include those on climate action (SDG 13) and nature-related goals to protect life on land and life in the water (SDGs 14 and 15). Social (S) goals include those dedicated to ensuring good health (SDG 3), quality education (SDG 4), and gender equality (SDG 5), among others. Economic goals include those for good jobs (SDG 8), innovation and infrastructure (SDG 9), and responsible consumption (SDG 12).See UN Sustainable Development Goals

ESG - Environmental, Social and Governance. Sustainability is the broadest concept, encompass-ing public and private action; ESG is typically used by the private sector to measure companies and screen investments.

Sustainability Targets - UN defined specific targets for SDG. For example, SDG 7 is affordable energey and its targets are to by 2030 substantially increase the share of renewable energy in the global energy mix, by 2030 double the global rate of improvement in energy efficiency among others.


TCFD- Task Force for Climate related Financial Disclosures. Widely accepted, being considered by Central Banks for their prudential supervision of climate risks at financial institutions.

SASB - Sustainability Accounting Standards Board was founded in 2011 to provide cross-comparable sustain-ability metrics. Defines key dimensions for climate change - Environment, Leadership & Governance, Business Model and Innovation, Soicial Capital and Human Capital. Gained prominence due to Blackrock requiring investee companies to disclose corporate performance in line with SASB metrics and recommendations

GRI -Global Reporting Inititiative started after Exxon Valdez disaster in 1997 with a set of guidelines and subsequently evolved into a disclosure framework with reporting standards.

WBCSD - The World Business Council for Sustain-able Development was formed after the UN Rio Summit in 1992, and it does research on corporate social responsibility and shares best practices on sustainability among its members.

PRI - Principles for Responsible Investing for Asset Managers.There are 6 Principles for Responsible Investment, these are a voluntary and aspirational set of investment principles and possible actions for incorporating ESG issues into investment practice.

PSI - Principles for Responsible Insurance for Insurance Companies. Launched in 2012, 4 Principles for Sustainable Insurance to guide better management of ESG issues and help insurance industry’s contribution to building a sustainable society.

PRB - Principles for Responsible Banking for Global Banks. 6 Principles for Responsible Banking for a sustainable banking system. Sustainability incorporated at the strategic, portfolio, and transactional levels, and across all business areas.


Climate Opportunity - Refers to the potential positive impacts related to climate change on an organization such as through resource efficiency and cost savings, the adoption and utilization of low-emission energy sources, the development of new products and services, and building resilience along the supply chain.

Climate Risk - Refers to the potential negative impacts of climate change on an organization. Physical risks emanating from climate change can be event-driven (acute) such as increased severity of extreme weather events (e.g., cyclones, droughts, floods, and fires). They can also relate to longer-term shifts (chronic) in precipitation and temperature and increased variability in weather patterns (e.g., sea level rise). Climate-related risks can also be associated with the transition to a lower-carbon global economy, the most common of which relate to policy and legal actions, technology changes, market responses, and reputational considerations.

Disclosure Areas - Decription of a company's Governance, Strategy, Risk Management, Metrics and Targets around climate risks and opportunities.

Disclosure Principles - Disclosure should be relevant, specific and complete, clear balanced and understandable, consistent format across time, comparable within a sector, industry or portfolio, reliable verifiable and objective and timely.