Sarvam AI (Axonwise Pvt Ltd), India’s leading AI startup, recently launched its newest LLM (large language model) named Sarvam-M has caused a debate about its approach and the problems with India’s AI (artificial intelligence).
Sarvam AI, an early company selected for the IndiaAI Mission, launched Sarvam-M. Based on Mistral Small, this 24-billion parameter open-source model is a key development for Indic AI. It supports 10 Indian languages, including Hindi, Bengali, Gujarati, Kannada, and Malayalam.
A limited 334 downloads in two days on Hugging Face led to considerable backlash for the project. Deedy Das, an investor at Menlo Ventures, openly criticised this model as “embarrassing”, and his comments, saying this small improvement wasn’t necessary, have sparked a heated debate among Indian AI experts.
Das highlighted the contrast with an open-source model created by two Korean college students, which achieved approximately 200,000 downloads.
Even though Sarvam promised more models later, their first one, Sarvam-M (which used a French AI model), faced criticism. Sarvam AI isn’t alone in getting few downloads. The government’s Param-1 model only had 12 downloads last week.
Das’s criticism isn’t just about downloads. He thinks Sarvam’s work shows they have the wrong goals. He said, “No one is asking for a slightly better 24B Indic model. “If you want to train models, there should be a very good reason for it.”
He added that Google and TWO.ai have models that are both cheaper and perform better in all these languages. He said, “I have nothing against Sarvam, but I just don’t think that at this moment their contributions are remotely commensurate to their funding”.
Sarvam has secured $41 million from investors like Lightspeed India Partners and Khosla Ventures, reaching a $111 million valuation by March 2025. Despite this, X users noted the model’s utility but also its need for improvement. Das believed Sarvam, despite their good Bulbul TTS model, needed to completely change how their systems work, pointing to DeepSeek in China as a model to follow.
Sarvam-M, despite low initial downloads, shows strong performance in Indian languages, outperforming some established models. While it gained criticism for its limited downloads, Sarvam AI emphasises its special role in improving AI for Indian languages and solving problems unique to India, like helping farmers.
The debate highlights the challenge and importance of building AI for India. We need AI that understands India’s many languages and helps its large population who speak those languages, instead of just aiming for global fame.
Indrajaal, an autonomous drone defence company, has introduced ‘Indrajaal Infra’, a new system built to protect critical infrastructure like nuclear plants, oil refineries, ports, airports, and power grids from drone threats.
The system uses the company’s SkyOS to automatically manage a multi-layered defence with sensors, spoofers, and jammers. This lets it watch and react to drones in real-time over huge areas, up to 4,000 square kilometres.
Indrajaal Infra is already in use at a naval port in Gujarat, with deployment initiated after successful field trials during recent cross-border incidents. Furthermore, a second system is being installed at India’s largest naval port in Karnataka.
Recent drone activity near the India-Pakistan border, though partly intercepted, highlighted the urgent need for better surveillance and defence, driving this launch.
Kiran Raju, Founder and CEO of Indrajaal, said, “Peacetime readiness is wartime insurance. The cost of protecting critical assets today is far lower than the cost of rebuilding them after an attack”.
The company highlighted how low-cost drone attacks can have significant consequences. Recent global incidents targeting oil terminals, energy infrastructure, and logistics hubs prove that drones can disrupt operations and cause economic and strategic damage.
Indrajaal Infra is an easy-to-use drone defence system that works with what’s already there and can protect different places. It uses smart tech (C5ISRT) to find and stop threats on its own. The company is working with defence groups, governments, and businesses to set it up in critical areas.
Niloufer Hospital in Hyderabad is now the first in the nation to utilise an AI-based diagnostic tool that performs rapid, needle-free blood tests.. This groundbreaking development for public health in India can eliminate the need for vials and laboratory wait times.
An app called Amruth Swasth Bharath, developed by health-tech startup Quick Vitals, utilises advanced face-scanning to give blood test results in 20 to 60 seconds, without a single drop of blood. It was recently revealed at the Red Hills campus of the Lakdikapul hospital.
Amruth Swasth Bharath utilises Photoplethysmography (PPG), a technology that works differently from regular blood tests. Instead of drawing blood, it uses light to check health through the skin. This allows it to assess key health indicators such as
Blood pressure
Oxygen saturation (SpO2)
Heart rate
Respiration rate
Heart rate variability (HRV)
Haemoglobin A1c
Stress levels
Pulse respiratory quotient (PRQ)
Sympathetic and parasympathetic nervous system activity.
This non-invasive method lets healthcare professionals do immediate health assessments with a smartphone or tablet camera. For long-term care, wearable sensors allow continuous monitoring.
Harish Bisam, founder of Quick Vitals, explained that their app makes health monitoring incredibly easy and compares it to simply taking a selfie. He said, “Our mobile face scanning system provides access to essential health data in under a minute. We believe this will bridge existing gaps in healthcare access, especially in underserved communities”.
Experts suggest this easy access could be vital for widespread health screenings, especially in rural and semi-urban areas where labs are scarce.
Amruth Swasth Bharath is set to significantly enhance maternal and child health programs by helping doctors quickly find conditions like anaemia that are often missed. Niloufer Hospital is already using it, and Maharashtra is next. Built by Quick Vitals, this AI-powered system keeps patient data private and secure. This technology will make good health information available to more people, cut down on testing delays, and reduce the need for uncomfortable tests.
Russia’s Sukhoi Su-57M fighter jet has completed its first flight using AI (artificial intelligence) assistance, a major breakthrough that could transform air combat in the future. This development has surprised US defence observers and created a strategic opening for India, which operates one of the world’s largest fleets of Russian-made Sukhoi aircraft.
Under the command of veteran test pilot Sergei Bogdan, the Su-57M recently completed its inaugural flight with significant AI assistance. While a human pilot remained onboard, an advanced AI system took charge of vital functions, including flight control, navigation, and target selection. This makes Russia one of the leading nations using AI in its fighter aircraft.
This AI is a key part of Russia’s long-running plan, the PAK FA program (started in 1999), to build advanced fighter jets. The Su-57M, an upgraded Su-57, has a powerful AL-51F-1 engine, advanced stealth, and long-range radar. Experts think it helps Russia compete better with American jets like the F-22 and F-35.
Although a pilot was present, AI took over much of the Su-57M’s flight, which will make pilots more efficient and quicker to react in dangerous situations by lightening their load. Defence experts predict this AI will be key to future air battles, enabling groups of fighters to respond to threats and missions with less human involvement.
Since India operates more than 270 Russian Su-30MKI aircraft and has strong defence ties with Russia, it’s well-positioned to team up on Russia’s AI-powered fighter jets. This could offer an affordable way for India to upgrade its current fleet with AI, which is crucial for today’s evolving air combat, especially after recent aerial encounters with Pakistan.
Google Cloud is significantly expanding its AI (artificial intelligence) infrastructure within India. This important step in a rapidly growing market means Google Cloud will now host its advanced AI models and offer its full set of AI tools locally.
Indian enterprises and developers are set to gain from Google’s AI advancements, starting with the locally hosted Gemini 1.5 Flash. This results in quicker response times and improved efficiency. Future yearly updates will include even more powerful models, such as Gemini 2.5 Flash, which will enhance India’s data sovereignty.
Google Cloud has two operational regions in India, Mumbai and Delhi NCR (serving regulated sectors like government), and is collaborating with the Indian government on the ₹100 billion IndiaAI Mission, which includes subsidised compute for startups and researchers.
Google Cloud’s AI technology is essential for government programs, like the iGOT Karmayogi platform. This platform uses Google’s advanced AI (like Vertex AI and special processors) and ready-made AI assistants for support, data analysis, and security to personalise training for civil servants.
This expansion not only emphasises India’s significance within Google Cloud’s global AI vision but also supports the government’s push for greater AI innovation and a stronger digital foundation.
During the four-day military conflict with Pakistan, starting on the night of May 6th-7th, Indian defence sources reported successfully striking several of the neighbouring country’s military sites.
India’s successful defence against Pakistani attacks was attributed to a powerful combination of three technologies. Space technology provided vital intelligence, advanced electronics offered strong defensive measures, and AI (artificial intelligence) delivered the analytical power needed for rapid and effective countermeasures.
According to sources, India’s possession of indigenous navigation, robust air defence with hard and soft kill options, and precise deep-penetration strike capabilities against Pakistan highlights the nation’s expanding strategic leverage through technological advancements.
A defence source said, “From detecting a radar picture of an enemy object in the sky, or taking a strategic position to shoot it down from land, sea and air was demonstrated using AI cloud-based state-of-the-art integrated air command and control systems”.
In last week’s Indo-Pak military conflict, India used a lot of its advanced technology to defend against Pakistani air threats. This was planned over five years ago, when the armed forces started looking at using AI in their strategies.
In 2018, the Defence Ministry formed the Defence AI Council (DAIC) and Defence AI Project Agency (DAIPA) after studying AI’s importance for security. By 2022, defence companies had a plan with 70 AI projects (40 done), and a total of 129 projects were approved by 2026 (77 done). In 2022, each branch of the military also set aside ₹100 crore for AI.
In the Western Theatre, the Indian Army now uses two AI systems from DRDO’s CAIR: the Intercept Management System (IMS) for automatic intelligence analysis and the Air Defence Control and Reporting System (ADC & RS) for detecting and stopping aerial threats early.
Kempegowda International Airport (KIA) operator BIAL (Bangalore International Airport Limited) have partnered with KPMG-India to leverage GenAI (Generative AI).
The BIAL said this strategic GenAI partnership with KPMG will significantly revolutionise airport operations, improve efficiency, and set new global aviation standards through adaptable and innovative AI models.
The partnership seeks to introduce a cutting-edge GenAI platform, which is uniquely designed for BIAL’s operational environment. This platform could really change the aviation industry, transforming how airports work and greatly improving customer experience.
This platform quickly handles and analyses huge amounts of live data using adaptable tech. It spots patterns and predicts trends, giving airports insights to prevent problems and keep things running smoothly.
By automating tasks, providing insights, and offering real-time AI support, this platform will make the airport more efficient, improve decisions, and increase passenger satisfaction.
The advanced platform will enhance airport operations with a consistent impact through strong data privacy, responsible AI, robust cybersecurity, and scalability.
George Fanthome, chief digital and information officer, BIAL, said, “At BIAL, we are committed to driving innovation at every level, and our association with KPMG in India is a significant step in unlocking the immense potential of GenAI at Kempegowda International Airport. GenAI’s capabilities, like real-time decision-making, predictive analytics, and adaptive intelligence, enhance our operations, leading to greater efficiency and sustainability”.
BIAL, already known for adopting cutting-edge technology, says its focus on data privacy, responsible AI, and scalability will make it a leading airport in the future.
Whether you are a student, a professional seeking a career transition, or an aspiring data scientist, you need to follow the right path to reach your desired goal. Which isn’t easy!
This is why we are here to help you with this blog, serving as your data science roadmap for 2025. Check it out and build a flourishing career in this thriving industry!
Data scientist roles are forecasted to grow by 36% between 2023 and 2033 globally, this pace is considerably quicker than the average for all occupations.
But!
What is the scope of data science in India?
The Indian data analytics market, valued at $3,551.8 million in 2024, is expected to surge to $21,286.4 million by 2030, driven by a strong growth trajectory.
This figure clearly illustrates that data science in India has a promising future. This powerful growth indicates an opportunity-filled and favourable career for aspiring data science professionals.
However, to carve a career in this field, you need a clear path that leads you to success. This is where the top data science AI course comes into play. This blog is your data science roadmap for 2025, guiding you through fundamental concepts, necessary tools, and more that you need to master this field. But first, let’s understand the basics of data science and AI (artificial intelligence)
Data Science & AI: Understanding the Basics
2025 is dramatically shaping the technological landscape with data science and AI (artificial intelligence). As the volume of data is exploding, the demand for skilled professionals is escalating rapidly. Before diving into these fields, you should understand what it exactly means.
What is Data Science?
Data science is a versatile field used to extract knowledge and valuable insights from both structured and unstructured data with the help of scientific methods, algorithms, processes, and systems.
Data scientists leverage various tools and techniques to:
Collect and clean data
Analyse data
Interpret data
Communicate findings
What is AI?
AI, or artificial intelligence, is a broader field which primarily focuses on creating computer programs and machines that are capable of performing various tasks that require human intelligence. This includes:
Learning
Reasoning
Problem-solving
Perception
Understanding language
Simply put, data science is all about extracting valuable information from sets of data, and AI is about building smart systems capable of performing human-like tasks.
Key Skills Required to Become a Data Scientist
You must have already seen the incredible impact of Data science and AI in 2025. This cutting-edge technology is driving innovation and valuable insights across numerous industries.
Let’s get down to the key skills you need to become a data scientist.
Domain Knowledge
Having a good understanding of the specific field or industry you will be working with, such as healthcare, marketing, finance, etc., is crucial. This knowledge helps you identify the right problems, understand the data limitations and context, and ensure that your solutions are actionable for businesses.
Computer Science/Programming Languages
This includes your programming abilities to handle, process, analyse, and model data using data science programming languages like Python, R, and tools like SQL. This also involves knowing computational concepts like data structures and algorithms, which are necessary for creating data pipelines and implementing models.
Mathematical Skills
Having a solid foundation in areas like Linear Algebra, Calculus, Probability, and Statistics is important. These skills provide you with a theoretical foundation for understanding how machine algorithms work, quantify uncertainty, interpret statistical results, and build deep knowledge of data and models.
Communication Skills
You should be able to clearly explain complex data findings and technical concepts easily to both non-technical and technical stakeholders, as well as your colleagues. Effective communication skills, including data storytelling and visualisation, help people understand, trust, and make decisions from your insights.
Learning Resources
You can acquire the aforementioned skills through many resources, including a data science AI course. Our full-stack data science with GenAI & ML course offers a structured curriculum with a good balance of theory and practical projects. You can acquire hands-on skills in this field and become an expert data scientist in 6 months.
Roadmap to Data Science & AI
To attain proficiency as a data scientist, you need a strong command of numerous disciplines, starting with foundational math and extending to advanced machine learning. The below-mentioned list is your data science roadmap that highlights key areas you need to explore in your journey to data science AI courses.
Mathematics
Knowledge of mathematics builds a bedrock for understanding algorithms and models in data science and AI.
Linear Algebra, Calculus, and Math Analysis: These are essential for understanding the core principles of various machine learning algorithms, particularly in optimisation and dimensionality reduction.
Differential Calculus: This is crucial for understanding techniques of optimisation used in training machine learning models.
Statistics
In-depth statistics knowledge provides tools for understanding data distribution, evaluating the performance of models, drawing inferences, etc.
Statistics & Central Limit Theorem: These are core concepts for understanding variability and creating statistical inference about populations from samples.
Probability & Sampling: Important for understanding the likelihood of events and how to pick good small groups from a big sample of data.
Hypothesis Testing: It is a fundamental and essential methodology in statistical inference and data analysis, which provides a structured approach to decision-making.
A/B Testing: It is a practical application of hypothesis testing, which is used for comparing different versions of the product or a feature.
Increasing Test Sensitivity: This is a technique to improve the power of AB testing to identify meaningful differences.
Econometrics
The econometrics field applies statistical methods to economic data, which offers key methods to analyse connections and predict future trends.
Fundamentals of Econometrics: This covers the main models and ideas to analyse economic data.
Regression: It is a powerful tool used for modelling the relationship between variables and making forecasts.
Time Series Analysis: These tools are used to understand data over time, which is crucial for predicting and identifying patterns.
Identifying Distributions: This technique is used for determining the probability distribution that represents a given set of data in the best way.
Coding
Having a strong foundation in programming is critical for data manipulation, analysis, and modelling.
Python Programming: Both experts and beginners prefer Python for data science as this is a dominant language in the AI and data science field due to its ease of use and extensive libraries.
Data Structures and Algorithms: These are core computer science concepts that help in building efficient solutions and handling data efficiently.
SQL Programming: This programming language is essential for interacting with databases and extracting meaningful data from them.
Exploratory Data Analysis
EDA, or exploratory data analysis, is another crucial step that includes understanding the data through statistical summaries and visualisation.
Data Understanding, Data Analysis and Visualisation: These are initial steps to understand the characteristics, quality, and insights from a set of data.
EDA with Python and Pandas: Using Python’s powerful Pandas library, you can do data manipulation and exploration.
EDA for Machine Learning: It focuses on insights that inform feature engineering and selection of the model.
EDA with Seaborn: Using the Seaborn library, you can create visually appealing and informative statistical graphics.
Machine Learning
ML, or machine learning, is the engine of AI. It empowers systems to learn patterns, make predictions, and enhance their performance, based on data.
Classic ML (Supervised, Unsupervised): This covers core algorithms for tasks such as classification, regression (supervised), clustering, and dimensionality reduction (unsupervised).
Advanced ML (Ensembles, Neural Networks): It explores more modern techniques, such as combining multiple models (ensembles) and introduces basic neural networks to improve performance.
Deep Learning
Deep learning is a branch of machine learning that uses AI neural networks with multiple layers to learn complex patterns from large datasets.
Fully connected networks: They form a basic structure of various neural network architectures.
Convolutional Neural Networks (CNN): These are designed to understand spatial relationships to effectively analyse image and video data.
Recurrent Neural Networks (RNN): These are designed to handle sequential information, which makes them ideal for tasks involving text, audio, and time-dependent data.
Long Short-Term Memory (LSTM): This is an advanced RNN that excels at capturing long-range dependencies in sequential data.
Transformers: The latest and highly effective architecture for NLP (natural language processing) and other types of data.
Transfer Learning: This technique uses knowledge from a pre-trained model to help a new model learn faster and better with less data on a similar task.
MLOps
MLOps, or machine learning operations, focuses on the practical factors of deploying and maintaining ML models in real-world applications.
Deployment Models: These are strategies and technologies used for creating trained models usable and accessible.
CI/CD (Continuous Integration/Continuous Deployment): It involves integrating automated processes to build, test, and deploy machine learning models.
Practical Application & Career Development
This step involves putting your theoretical knowledge into practice, which is crucial in a data science and AI course.
Importance of Hands-on Projects: To solidify your understanding, you need to apply your knowledge to real-world problems. This also helps in building a strong portfolio.
Building a Portfolio: You need to showcase your skills in data science and AI through a solid portfolio.
Competitions and Community Engagement: You should engage in data science and AI competitions to acquire practical experience. Engage yourself in the community to discover new tools and learn from experienced professionals. This will help you grow your practical skillset. You can also enrol on our data science course in Kolkata to get hands-on training in the field and acquire job-ready skills.
In Summary
The data science and AI field offers numerous career opportunities in 2025 and beyond.
From foundational knowledge to advanced techniques and practical application, this data science roadmap outlined the critical steps that will greatly help you in your career.
The present is data-driven, and the future is expected to be more analytics-driven. With a focused approach and consistent hard work, you can become a data scientist with expertise in AI.
Our 6-month blended program can help you pave your way to becoming an expert data scientist.
So, don’t wait anymore. Contact us today: 9836 423 755/ 6289 562 294 to take the first right step towards a lucrative data science career!
FAQs
What is the future of data science in 2025?
Data science is expected to be leveraged and implemented in more industries with more focus on specialised applications like real-time analytics and explainable AI. The need for skilled data scientists will remain significant.
What is the scope of AI in 2025?
AI is projected to see wider adoption in automation, solving complex challenges, personalised experiences, and intelligent systems. You will see advancements in fields like computer vision, NLP, robotics, etc.
What is the roadmap for data science AI courses?
Data science roadmap for AI courses begins with foundational mathematics, programming, statistics, and advances to data analysis, deep learning, machine learning, and specialised AI applications. Remember, working on practical projects is crucial to building a strong foundation in these fields.
Will AI replace data science jobs?
AI is expected to automate various human tasks, but it is likely to augment their roles, not replace them. Data scientists will remain in demand for solving complex problems and extracting strategic insights.
Which is the best artificial intelligence and data science course?
The best artificial intelligence and data science course for your career depends on your goals, prior knowledge, and educational background. You need to look for courses that offer industry-relevant projects, a comprehensive curriculum, career support, and experienced instructors like Karmick Institute’s data science with GenAI & ML program.
How to learn AI in 2025?
You can start with learning a programming language like Python, statistics, and mathematics. After that, move to AI fundamentals, machine learning, deep learning, hands-on projects, and consistent practice.
Which is the most in-demand AI career?
Roles like machine learning engineer, natural language processing engineer, AI scientist, and computer vision engineer are highly demanded across various domains.
Is data science worth in 2025?
Absolutely! Data science is worth it in 2025 and beyond, driven by its growing applications. Due to the increasing volume of data, there will be a strong demand for data scientists across various sectors in 2025 and in the coming future.
What skills does a data scientist need?
Data scientists need to possess essential skills, including programming (Python), machine learning, statistics, data visualisation, data wrangling, communication, and domain knowledge.
How long does it take to become a data scientist?
Becoming a data scientist depends on several factors, including educational background, prior knowledge, grasping power, mode of learning, data science AI course curriculum, etc. However, you can expect to master this field between 6 months to a few years.
BAT VC, an early-stage venture capital firm headquartered in New York. This company has revealed its intention to invest a whopping $100 million through its latest fund. Their main focus will be on Indian startups in the AI (artificial intelligence) and deep technology sectors, with a specific interest in fintech and B2B SaaS (software-as-a-service) companies.
A newly launched India-focused initiative from BAT VC will be under the leadership of three general partners, including Manish Maheshwari, who previously headed X India, formerly known as Twitter. Aditya Mishra and Ravi Metta are the other two partners.
BAT VC says India’s AI is booming, calling it a “golden era.” It’s growing fast at 32% each year and is expected to hit $23 billion by 2027. Enterprise software (SaaS) for businesses is also doing great, growing even faster at 35% yearly and is bigger than the global average.
The VC firm highlighted that investing in AI companies with connections between the US and India is very promising. This area saw a huge increase to $4.7 billion in 2023 because of the talented people, available funding, and access to markets in both countries.
Manish Maheshwari, General Partner and India Head at BAT VC, said, “My move to Bengaluru underscores our conviction in India’s potential to lead the next wave of AI-driven global growth”.
The VC firm believes that the strong interest in Fund II from big investors in the US and India shows that it is now a central investment for global players, not just a developing market.
Ravi Metta, General Partner, BAT VC, said, “Our technical depth enables us to identify high-impact AI startups early and guide them through global scale-up”.
Monday 12, May 2025: According to a report, close to 50% of Indian tech professionals receive AI (artificial intelligence) skilling support from their companies.
The job portal Naukri conducted a survey on National Technology Day, which highlighted a significant trend in India’s tech ecosystem.
With responses from more than 16,000 technology professionals across various sectors, the survey indicates a clear shift towards planned upskilling in addition to continuous self-learning.
The report indicates that AI upskilling is now mainstream among Indian professionals at all levels.
In major tech cities like Bengaluru and Gurugram, over 50% of the new graduates surveyed said they either know some basic AI or are fully involved in AI training.
In some cities, about 33% of workers are being actively trained in AI by their companies. IT/software workers get more training because their field changes fast and requires it more.
For professionals with 10-15 years of experience, GenAI was the top skill focus for 42%, increasing to 46% for those with over 15 years of experience. At the same time, many young professionals are also interested in learning about Cloud computing, DevOps, and Data Engineering. This shows that there’s a general interest in keeping up with new technologies.
Apart from job losses (worrying 18% of tech workers), the survey found bigger concerns: not enough time to learn new skills, using old tech, and small pay raises. Specifically, 20% said they don’t have enough to dedicate to upskilling.
The report also highlighted that outdated tech stacks were a significant obstacle for 17% of respondents. Furthermore, 46% of technology professionals expressed major concerns regarding low salary increments.
info@karmickinstitute.com
+919836423755 
+919836423755
Login
