I am interested in machine learning methods and theory for genomics, focusing on modeling and
mechanistic interpretation of structural, epigenetic, and regulatory variation. My work spans
structural variant detection in cancer, nanopore raw signal analysis, behavioral detection in fish, copy number variation in
insect genomes, and host–pathogen evolutionary dynamics.
I am part of the Gemstone Honors Program and the Computer Science and Entomology Departmental
Honors Programs at Maryland, where I hold a 3.8 GPA. I am look for PhD programs in
computational biology and machine learning.
Research
Raw Signal Analysis Methods
Student Research Assistant, Firtina Lab, University of Maryland ·
October 2025 – present
Building methods for nanopore sequencing raw signal analysis: methylation detection, error
correction, and assembly. This work includes RawMod, a modification-agnostic classifier that
detects DNA modifications without requiring a control sample. RawMod pileups raw signal from
up to 30 reads at a candidate site, refines each read against a reference k-mer lookup table,
and passes the pileup through a convolutional trunk feeding a transformer that attends across
reads to reach a consensus modification likelihood.
RawMod architecture: signal pileup, per-read convolutional trunk, and cross-read transformer consensus.
On held-out ONT and UMCES benchmarks, RawMod reached F1 scores of 0.58 and 0.40 — 1.1x
and 3.08x Dorado's, respectively — and generalized to four modification types withheld
during training at ≥3.6x Dorado's F1, while Dorado misidentified the modification type on
every held-out call. RawMod won the Best Poster Award at HiTSeq, ISMB 2026.
Student Research Assistant and Mentor Liaison, Gemstone Program, University of Maryland
· August 2023 – present
Developing a computer vision classification model to identify the social behaviors of
Lake Malawi cichlids. I lead a team of 11 students in the Gemstone Honors Program
under the mentorship of Dr. Abhinav Shrivastava in collaboration with Dr. Scott Juntti.
Segmentation and keypoint tracking on tank footage.
Student Intern, Kolmogorov Lab, National Cancer Institute · May 2026 – present
Developing a deep learning method to identify and localize complex structural
variation in cancer genomes. The tool combines copy-number segments and breakpoint
graphs from long read and short read callers to detect BFB, chromothripsis, ecDNA,
and seismic amplification events.
Computational Biology Research Assistant, Fritz Lab, University of Maryland ·
February 2024 – present
Analyzing the basis of Bt pesticide resistance and copy number variation in Helicoverpa zea.
This work includes analyzing WGS short read data, structural bioinformatics analysis, and
benchmarking short read SV callers with the Molloy Lab.
CNV frequency across sampled populations, and a structural model of the toxin–trypsin interface.
Student Research Assistant, Bruns Lab, University of Maryland ·
January 2024 – present
Building ODE models to simulate host–pathogen coevolution and study temporal adaptation
and allelic variation, incorporating seasonal forcing and birth pulses.
Evolutionary Stable Strategy comparison across different species.
Bioinformatics Intern, AstraZeneca, Vaccines & Immune Therapies ·
May – August 2025
Engineered and deployed a bacterial genome-processing pipeline supporting vaccine R&D
surveillance efforts, and developed an unsupervised learning approach to identify
non-typeable Haemophilus influenzae cell-type clusters.
US Department of Agriculture – Agricultural Research Service
Biological Science Aid, Beltsville, MD · April 2024 – January 2025
Applied plant breeding methods to generate improved strawberry cultivars emphasizing disease
resistance, and deployed computer vision models for real-time detection of diseased plants
and stray animals using autonomous robots.
Best Poster Award, HiTSeq, ISMB, July 2026 — RawMod: A Deep Learning
Approach to Modification Agnostic Classification in Nanopore Sequencing.
Best Poster Award, Entomological Society of Washington Student Showcase,
May 2026 — Correlations between Read Depth, Copy Number Variation, and Bt Resistance in
Helicoverpa zea.
2nd Prize, Crop Science Society of America Poster Competition, PAG33: Plant
and Animal Genome Conference, January 2026 — Correlations between Read Depth, Copy
Number Variation, and Bt Resistance in Helicoverpa zea.
Best Talk Award, University of Maryland Gemstone Library Awards, April 2025
— Investigating the Link between the Ora4 Gene and Social Behaviors in African Cichlids
through the Use of Computer Vision. Awarded $2,000 in research funding.
Honors College Research Grant, CS Professional Development Fund, GO Fund, and
Ernest N. Cory Undergraduate Scholarship — funds collectively totaling over
$5,000 in support of independent research and conference travel.
Papers
CRANE: Correcting Errors in Raw Nanopore Signals Using Hidden Markov Models.
Simon Ambrozak, Ulysse McConnell, Bhargav Srinivasan, Burak Ozkan,
Ernest Zhang, Can Firtina. arXiv, 2026.
Artificial Intelligence and Pain Management: Cautiously Optimistic. Bhargav Srinivasan, Archana Venkataraman, Srinivasa N. Raja. Editorial, Pain Management, September 2024.
RawMod: A Deep Learning Approach to Modification Agnostic Classification in Nanopore
Sequencing. Bhargav Srinivasan, Simon Ambrozak, Can Firtina. In preparation.
FSHFormer: A Computer Vision Model for Automated Cichlid Behavior Annotation. Bhargav Srinivasan, William Lamousin, Charles Phan, Jason Liu, Shubh Sharma,
Vijay Jayasuriya, Ariel Bazan, Cade McGeehan, Ofure Osunbor, Tahir Haroon, Pulkit Kumar,
Abhinav Shrivastava. In preparation.
When Should a Pathogen Strike? Navigating the Landscape of Birth Pulses and Mating
Season. Bhargav Srinivasan, Sam Hulse, Emily Bruns. In preparation.
Talks
A Deep Learning Approach to Identifying Complex Structural Variation in the Genome. Invited seminar talk, Cancer Data Science Laboratory, National Cancer
Institute, Bethesda, MD. July 2026. Related poster (PDF)
Unsupervised Learning for DNA Modification Profiles from Nanopore Sequencing Data. Invited talk, Algorithmic Bioinformatics Seminar, University of Maryland,
College Park, MD. April 2026.
Developing a Molecular Surveillance Dashboard. Invited talk, AstraZeneca Vaccines & Immune Therapies Intern
Symposium, Gaithersburg, MD. August 2025.
Host Life History vs. Pathogen Optimization. Invited talk, Quantitative Ecological & Evolutionary Dynamics
Seminar, University of Maryland, College Park, MD. April 2025.Only undergraduate speaker in the seminar's history.
Posters
RawMod: A Deep Learning Approach to Modification Agnostic Classification in Nanopore
Sequencing. ISMB: Intelligent Systems for Molecular Biology, Washington, D.C.
July 2026.HiTSeq Best Poster Award. Poster (PDF)
A Deep Learning Approach to Identifying Complex Structural Variation in the Genome. Poster presentation and flash talk, NCI Summer School on Algorithmic
Cancer Biology, Bethesda, MD. July 2026. Poster (PDF)
A Machine Learning Approach to Identify Unique Non-typeable Haemophilus
influenzae (NTHi) Clusters Using Lipooligosaccharide (LOS) Biosynthetic Genes. ASM Microbe 2026, Washington, D.C. June 2026.
Correlations between Read Depth, Copy Number Variation, and Bt Resistance in
Helicoverpa zea. Entomological Society of Washington Student Showcase, Washington, D.C.
May 2026.Best Poster Award. Poster (PDF)
Correlations between Read Depth, Copy Number Variation, and Bt Resistance in
Helicoverpa zea. PAG33: Plant and Animal Genome Conference, San Diego, CA.
January 2026.2nd Prize, Crop Science Society of America Poster Competition. Poster (PDF)
When Should a Pathogen Strike: Navigating the Landscape of Birth Pulses and Mating
Season. Ecology and Evolution of Infectious Diseases North-East Meeting,
University Park, PA. November 2025.
Investigating the Link between the Ora4 Gene and Social Behaviors in African Cichlids
through the Use of Computer Vision. University of Maryland Gemstone Library Awards, College Park, MD.
April 2025.Best Talk Award.
Software
Main contributor. All tools are actively under development.
