Pratham Sahu

Fourth Year Undergraduate spratham21@iitk.ac.in | ‡ Prathamsahu52

Department of Computer Science and Engineering, IIT Kanpur Œ Website | Ó +91-7619678791| ¯ Pratham Sahu

Academic Qualifications
Year Degree/Certificate Institute CPI/%
2021 - Present B.Tech Indian Institute of Technology Kanpur 9.31/10
2021 Karnataka State Board(XII) St Jerome’s PU College, Bangalore 98.3%
2019 ICSE(X) Vibgyor High, Bangalore 96.4%

Scholastic Achievements
• Secured All India Rank 131 in JEE Advanced 2021, conducted by IIT Kharagpur, among 1,50,000 shortlisted candidates.
• Secured an All India Rank of 87 in JEE Mains 2021, conducted by NTA among 1.1 million candidates.
• Secured AIR 44 in Indian National Physics Olympiad(INPhO), in 2021 and made it to National Selection Camp(IPhO).
• Secured AIR 31 Indian National Astronomy Olympiad(IAO), in 2021 and made it to National Selection Camp(IOAA).
• Secured AIR 19 Indian National Chemistry Olympiad(INChO), in 2021 and made it to National Selection Camp(IChO).
• Recipient of the Directors Scholarship, IIT Kanpur in the year 2022 for having an exceptional JEE Advanced rank.
• Awarded KVPY SA 2019 fellowship, securing an All India Rank 340 conducted by Indian Institute of Science, Bangalore.
• Awarded KVPY SX 2020 fellowship, securing an All India Rank 148 conducted by Indian Institute of Science, Bangalore.
• Received the Academic Excellence Award for exceptional academic performance in 2021-22 and 2022-2023 academic session.
• Recipient of the National Talent Search Examination(NTSE) Scholarship conducted by NCERT in 2019.
Internships and Key Projects
• Adobe Research, Bangalore (May ’24 - current)
Intern Research Assistant, Big Data Experience Lab
– Working as a part of the Data-driven Systems, Insights, Experiences team.
– Working on data and domain aware strategy generation by leveraging customer data and finetuning of LLMs.
• Yonsei Vision and Learning Laboratory, Seoul, South Korea | Prof. Jonghyun Choi (May ’23 - Sep ’23 )
Intern Research Assistant
– Engaged in research on developing efficient deep neural architectures and effective continual machine learning algorithms.
– Performed extensive survey on sampling, improving time and efficiency with limited accuracy loss of ML algorithms.
– Assessed on developing efficient coresets for streaming data to mitigate forgetting in incremental deep learning setups.
– Worked on incremental classification to come up with a novel technique to evolve existing methods in episodic-replay.
• MSenseAI, Bengaluru | Platform Software Engineering Intern (Jul ’22 - Aug ’22 )
– Implemented open source IoT Platform ThingsBoard to add functionalities to suit specific needs of the product.
– Successfully implemented open source JS library Annotorious to implement picture annotating on the platform.
Competitions and Workshops
• ISC Student Cluster Competition
Prof. Preeti Malakar & Prof. Swarnendu Biswas (Aug ’23 - May ’24 )
– Led a team of six third-year students in the world’s largest student HPC Competition, becoming the first Indian
team to get selected. Finished 8th out of 21 shortlisted teams, being the only new team in the competition.
– Optimized the µphys package in ICON application, using OpenACC directives to make the code portable to CUDA
backends, and employed parallelization techniques to achieve a 20x improvement in performance.
– Built, ran, visualized, and profiled the performance of the NekoCFD application on CPU and GPU backends on the
Bridges2 supercomputer. Demonstrated weak and strong scaling performance and visualised using paraviewCLI.
Selected Projects
• Controlled Interthread memory sharing in multi-threaded applications
Linux Kernel Programming| Prof. Debadatta Mishra| Report (Jan ’24 - May ’24 )
– Designed a novel memory framework for threads to achieve previlege seperation in the same address space.
– Used TLB entries to boost previleges temporarily on cores with threads having higher previleges for the same page. Used
inter-processor interrupts to stop other cores from accessing boosted Page Table entries and thus ensuring correctness.
Optimized for efficiency using TSX provided by a few older x86 systems and pointed out issues with same.
– Implemented the design on the Linux kernel using modules, core kernel code, and provided user-space APIs.
– Conducted thorough testing and benchmarking, verifying correctness and assessing memory access time trade-offs.
• Modelling Performance Variability in HPC Clusters(Undergraduate Project)
Ongoing Prof. Preeti Malakar |Report (Jan ’24 - current)
– Analyzed job interference impact on performance variability in a large-scale production supercomputer(Param Sanganak).
– Characterized supercomputer jobs using profilers(IPMPI), I/O tracing, network tracing, and hardware counters(perf ).
– Designed algorithm to mitigate job interference effects on performance, integrated and simulated on Slurm Scheduler.
– Received an A* for exceptional contributions to research in the field of performance and variability modelling.
 • PuppyLove2.0 | Programming Club ‡ (Jan ’23 - Apr ’23 )
– Built a cryptographically secure dating application for the campus community which ensured zero-server side knowledge.
– Deployed the application using Kubernetes along with security measures to ward off large DOS attacks on the server.
– Achieved successful registration of 2800 campus residents and around 300 matches being made by our application.
• Python Compiler | CS335 Course Project ‡| Prof. Swarnendu Biswas (Jan ’24 - Apr ’24 )
– Developed an end-to-end compiler to convert a subset of Python language features to x86 assembly code.
– Designed lexer using flex, parser using bison and semantic analyzer to support for classes, multilevel inheritance, arrays,
and non-primitive types in python. Included support for type checking and implicit type conversion.
– Used a hierarchical symbol table and Abstract Syntax tree based IR to form a 3AC IR, which was then converted
to x86 assembly code using template functions. Awared 97% marks for complete implementation.
• Operating System Design | CS330 Course Project ‡| Prof. Swarnendu Biswas (Aug ’23 - Nov ’23 )
– Implemented support for system call tracing and function call tracing on a toy OS with x86 architecture.
– Implemented mmap, munmap and remap semantics in the OS ensuring least fragmentation of virtual memory.
– Implemented regular(copy all) and Copy-on-Write(CoW) semantic for fork handling within the toy OS.
• Scalable Parallel Feature Extraction and Tracking for Large Time-varying 3D Volume Data
CS677 Course Project ‡| Prof. Preeti Malakar & Prof. Soumya Dutta (Aug ’23 - Nov ’23 )
– Implemented a high performance parallel feature extraction and tracking algorithm for large 3D volume data.
– Validated the algorithm on large scale datasets and compared the performance with existing state-of-the-art methods.
– Performed weak-scaling and strong-scaling analysis to evaluate the performance of the algorithm on large scale clusters.
• CSE-Bubble | CS220 Course Project ‡| Prof. Urbi Chatterjee (Jan ’23 - Apr ’23 )
– Implemented a Verilog hardware description of a 32-bit processor, featuring ISA, ALU, and a memory unit.
– Verified correctness by running RISC-V(MPI) assembly code for Bubble Sort on the on a FPGA board.
• CampusPay | CS253 Course Project ‡| Prof. Indranil Saha (Jan ’23 - Apr ’23 )
– Developed the code-base for a website to handle finances and dues for ease of campus community and vendors.
– Utilised Django framework to create the webite’s backend, ReactJS for the website’s frontend and SQLite for database.
Relevant Courses
Computer Architecture Parallel Computing(A*) Linux Kernel Programming
Compilers Computer Organisation Networks
CUDA programming Analysis and Design of Algorithms Data Structtures and Algorithms
Mathematical Logic Fundamentals of Computing Undergraduate Project(A*)
Operating Systems Large data analysis and visualisation(A*) Theory of Computation
Probability Introduction to Machine Learning Software Development and Operations

Technical Skills
• Programming Languages: C, C++, Python, Java, Javascript, Solidity, RUST, CUDA, DPC++
• Software and Libraries: gdb, perf, Tau profiler, Nvidia Nsight, Numpy, Pandas, MatplotLib, MERN stack, Git, NextJS,
PyTorch, Django, Kubernetes, Docker
• Exposure: Bash, Linux Kernel Programming, VerilogHDL, MIPS, XHR, AJAX, MQTT, GCP
Teaching
• Tutor, Data Structures and Algorithms(ESO207) (July ’24 - current)
Volunteering
• Coordinator, Programming Club IITK (Mar ’23 - Apr ’24 )
– Managed one of the most active clubs of IIT Kanpur which delves into the multiple domains of programming.
– Worked on open source projects targetting the campus community such as StudentSearch and PuppyLove.
– Organised Linux install fest, Spring Camp (2 weeks of workshops across 5 domains) and Opportunity Open Source
conference with Canonical and Linux Foundation for the benefit of the Campus Programming Community.
– Mentored 1̃5 students to get introduced to the concepts of Operating Systems and understanding the Linux Kernel.
• Secretary, Programming Club IITK (Sep ’22 - Mar ’23 )
– Assisted in conducting competitions, activities, lectures, and workshops for programming enthusiasts in IIT Kanpur.
• Student Guide, Counselling Service IIT Kanpur (Sep ’22 - May ’23 )
– Mentored a group of 6 freshmen academically and emotionally to get acclimatized to the new college environment.