Best Place To BuildProf. Shweta Agrawal, CSE | "Real-life cryptography is cooler than Imitation Game-the movie"| Ep.13
CHAPTERS
- 0:00 – 0:38
Cryptography as the art of secrets—and a balance of structure vs randomness
Shweta frames cryptography as “the art of keeping secrets” and hints at the core intellectual tension: ciphertext must look random, yet hide enough structure to enable legitimate use. The chapter sets the tone that modern cryptography is broader and “cooler” than pop-culture portrayals.
- •Cryptography’s core goal: enabling secrecy and trust in communication
- •The central theme: a fine dance between structure and randomness
- •Cryptography as an intellectual/artistic pursuit (not just engineering)
- 0:38 – 1:34
Podcast setup: meeting Prof. Shweta Agrawal at IIT Madras
The host introduces the series premise—builders at IIT Madras—and brings in Prof. Shweta Agrawal as the guest. Her background, chair position, and awards establish her as a leading cryptography researcher.
- •Context: Best Place to Build podcast at IIT Madras innovation hub
- •Guest introduction: Prof. Shweta Agrawal (CSE), awards and chair position
- •Cryptography positioned as the episode’s focus
- 1:34 – 2:18
What cryptography actually does: privacy, authentication, and more
Shweta gives a practical definition of cryptography and expands it beyond “hiding messages.” She explains multiple cryptographic goals, including private communication and proving identity/authenticity.
- •Cryptography enables secret communication and secure interactions
- •Goals include privacy and authentication/identity verification
- •Cryptography as the science of achieving these guarantees
- 2:18 – 4:33
From Caesar cipher to modern cryptography: proofs and hard problems
Starting with ancient ciphers (e.g., Caesar cipher) and wartime codebreaking, Shweta explains what modern cryptography changed: relying on mathematical reduction to well-studied hard problems. The emphasis moves from “clever obscurity” to provable security assumptions.
- •Historical roots: Caesar cipher and long-standing human need for secrecy
- •Wartime lessons: systems can fail if designers “miss something”
- •Modern approach: reduce breaking security to solving a hard math problem
- •Greater confidence via decades of scrutiny of underlying assumptions
- 4:33 – 7:05
Everyday and business use-cases: shopping security and private joint computation
The discussion shifts to why cryptography matters outside national security. Shweta explains online payments as a basic example, then introduces multiparty computation as a way for entities to compute shared outcomes without revealing private inputs.
- •Online transactions require protection against network eavesdroppers
- •Attack scenario: intercepted card data during transmission
- •Business example: companies evaluating mergers without revealing assets
- •Multiparty computation: learn only the output, nothing else about inputs
- 7:05 – 7:59
Threat models: eavesdroppers vs active attackers (as algorithms)
Shweta clarifies the language cryptographers use: attackers are modeled as algorithms, not specific people. She distinguishes passive eavesdropping from stronger adversaries who can also modify communications.
- •In cryptography, “attacker” is an algorithmic model
- •Eavesdropper = passive listener on the channel
- •Active attackers may alter messages or manipulate the channel
- •Security must account for multiple attack capabilities
- 7:59 – 9:15
What “hard” means in CS: efficient (polynomial-time) vs infeasible computation
Shweta defines computational hardness in the formal computer-science sense: no efficient probabilistic algorithm should solve the underlying problem. She connects hardness to realistic time horizons—from years to the age of humanity—through adjustable security parameters.
- •Hard = no efficient probabilistic algorithm can solve it
- •Efficiency modeled as polynomial-time computation
- •Security levels depend on the acceptable attack timeframe
- •Hardness parameters can be tuned to application needs
- 9:15 – 12:32
Algorithms and RSA: public-key encryption and the role of secret keys
The host asks what an algorithm is, and Shweta bridges from simple examples to real cryptosystems like RSA. She explains RSA’s public/private key split: anyone can encrypt with the public key, but only the secret key enables decryption.
- •Algorithm = step-by-step procedure mapping input to output
- •RSA as a foundational, influential cryptosystem
- •Ciphertext should be infeasible to invert without solving a hard problem
- •Public-key encryption: public key for encryption, secret key for decryption
- •Practical takeaway: protect the secret key; public key can be shared
- 12:32 – 17:14
P vs NP and the quantum threat: why RSA breaks under quantum algorithms
Shweta introduces P vs NP as the central open problem behind hardness intuition. She then explains that quantum algorithms can change what’s considered hard—famously making RSA easy to break in principle—so cryptography must plan for “post-quantum” security now.
- •P: efficiently solvable problems; NP: efficiently verifiable solutions
- •Cryptography needs problems believed not to be efficiently solvable
- •Algorithms can move problems into “easy” territory over time
- •Quantum setting differs; some classically hard problems become easy
- •RSA is believed hard classically but easy for quantum algorithms
- 17:14 – 24:58
Research focus: computing on encrypted data and the structure–randomness paradox
Shweta describes her theoretical cryptography work, focusing on enabling meaningful computation on encrypted inputs (motivated by ML and sensitive datasets like genomics). She explains the paradox: encryption should hide everything (look like noise) yet retain hidden structure enabling computation.
- •Beyond privacy: encryption that still supports computation
- •Motivation: ML/medical research on sensitive genomic data
- •Ciphertexts must be indistinguishable (random-looking) to attackers
- •Yet computation (add/multiply/decrypt) needs hidden structure
- •Core idea: cryptography as balancing “form and formlessness”
- 24:58 – 28:02
Functional encryption, access control, and lattices for post-quantum cryptography
The conversation names the concept explicitly: functional encryption, where keys authorize specific computations on ciphertexts. Shweta also covers attribute-based encryption as a policy/identity-driven special case and introduces lattice-based cryptography as a rich foundation for post-quantum secure schemes and new capabilities.
- •Functional encryption: learn only an approved function’s output from encrypted data
- •Authorization via keys: only correct/allowed computations can be decrypted
- •Attribute-based cryptography: decrypt if attributes satisfy an access policy
- •Lattice-based cryptography: hard lattice problems replace factoring/discrete log
- •Lattices enable both post-quantum security and new cryptographic power
- 28:02 – 29:41
Hard problems: factoring intuition and how hardness differs from engineering difficulty
Asked to compare to “hard” engineering equations, Shweta grounds hardness in algorithmic scaling. Using factoring, she explains why brute force becomes exponential in input size and why cryptography relies on the belief that no dramatically faster classical algorithm exists.
- •Example hard problem: factoring product of two large primes
- •Brute force factoring scales exponentially in input length (bits)
- •There are improvements over brute force, but not order-changing classically
- •Cryptographic confidence comes from long-term research scrutiny
- 29:41 – 34:23
What computer science is (and isn’t): computing, complexity, systems, and ML
Shweta pushes back on the notion that CS equals coding, using Dijkstra’s quote to emphasize “computing” over “computers.” She outlines major subfields—from theory/complexity to systems and machine learning—and positions programming as a tool rather than the definition of the field.
- •CS is about computing: what can be computed and at what cost
- •Dijkstra quote: “computer science” vs “knife science” analogy
- •Theory: complexity, hardness, reductions, cryptography
- •Systems: OS, compilers, networks, programming languages
- •ML/AI: questions about learning, thinking, creativity, and scale
- 34:23 – 39:42
Choosing CS wisely: interest-driven study, student pressure, and exploring beyond the track
Shweta argues that students should ask ‘why this field?’ rather than defaulting to job-driven choices. She discusses the pressure students face, the importance of exploration (electives and broad exposure), and the reality that informed decisions are still difficult in India’s university ecosystem.
- •Career choices should be informed by personal interests and exposure
- •CS is young with many open problems and rapid discovery
- •Students experience intense pressure; some feel mismatched to the field
- •Exploration (electives, breadth) can reveal better fits
- •Ideal: stronger university ecosystems enabling broader undergraduate exploration
- 39:42 – 46:31
India’s rise in cryptography and building CyStar as a cybersecurity center
Shweta describes India’s rapid growth in public-key cryptography over the past decade, including stronger publication presence and talent development. She then explains CyStar’s mission and structure: a cybersecurity center spanning theory, systems/applied security, and bridging real-world models with formal methods.
- •10-year shift: from minimal public-key crypto presence to strong representation
- •Growth driven by returning researchers and training new students locally
- •CyStar = Security, Cryptography, Trust, and Reliability center
- •Leadership split: applied leakage/implementation, theory/algorithms, bridging models
- •Activities: outreach, hackathons, schools, incubators, supporting PhD talent
- 46:31 – 51:22
Women in STEM: pipeline thinning, bias in leadership perceptions, and performance realities
Shweta differentiates India’s challenges from those in the US, noting the pipeline narrows at higher levels even when women perform strongly. She cites studies illustrating leadership perception bias and highlights that women students can excel academically while still facing structural and cultural barriers.
- •In India, undergrad participation can be healthier than stereotypes suggest
- •Major issue: ‘thinning out’ as careers progress toward leadership roles
- •Leadership bias example: ‘head of the table’ perception experiment
- •Performance note: women students often have higher average CGPAs in IIT studies
- •Residential/coaching ecosystem and social factors affect participation
- 51:22 – 1:03:12
Personal journey, returning to India, and Atma Shraddha (self-belief)
Shweta recounts a less scripted path—exploration across art, math, physics, and CS—leading to cryptography by chance. She explains returning to India as a deeply personal commitment to giving back, then argues the country needs stronger Atma Shraddha: confidence to build world-class work locally based on informed choices, not inherited perceptions.
- •Education as exploration: broad learning and discovering cryptography serendipitously
- •Decision to return was pre-committed; motivated by scarcity of opportunities observed growing up
- •Building domestic capability: contributing to India’s research ecosystem
- •Atma Shraddha: overcoming post-colonial ‘abroad = better’ reflex
- •Key nuance: going abroad can be right—but should be an informed, evidence-based choice
- 1:03:12 – 1:05:44
Art, abstract expressionism, and closing thoughts on beauty across disciplines
In the closing, Shweta connects cryptography back to art—especially abstract expressionism—as another structure–randomness dialogue. She argues disciplinary boundaries are often artificial and that a shared pursuit of beauty links math, art, music, and the humanities.
- •Shweta continues doing art; sees cryptography itself as an art form
- •Abstract expressionism analogy: form vs formlessness mirrors cryptography
- •Communication vs concealment as a thematic parallel
- •Beauty as a unifying driver across fields
- •Podcast wrap-up and call to explore her lectures/work
