I'm a biotechnology researcher who builds software. I take a biological question, turn it into a computational workflow, and try to make every step of that workflow something another researcher can inspect, test and improve.
I'm doing an MSc in Biotechnology (Industrial Biotechnology) at Gujarat Biotechnology University. My work so far covers CRISPR guide design, machine learning for fermentation, structural bioinformatics and, right now, the search for phage enzymes that kill drug-resistant bacteria.
Endolysins against ESKAPE pathogens - my MSc dissertation (Oct 2026 - Jun 2027, supervised by Dr. Dhaval Patel)
Endolysins are enzymes that bacteriophages use to break open bacterial cell walls, which makes them a promising alternative to antibiotics. I'm mining public metagenomic datasets for new ones and ranking the best candidates against six drug-resistant hospital pathogens.
- Find - build a curated, non-redundant set of phage proteins from metagenomes
- Detect - pick out endolysins with protein language models, checked against domain search
- Match - predict which bacteria each endolysin's phage infects
- Rank - shortlist 10 to 20 candidates using sequence, structure and docking evidence
Scan a sequence, find every NGG site, and surface the 20-nucleotide guide beside it. The real tools add ranking, specificity checks and a record of where every sequence came from.
|
Finds one SpCas9 guide that targets several homologous plant genes, with mismatch-aware consensus design, per-gene evidence and explicit PASS / REVIEW / FAIL validation.
|
A careful guide-design workbench for OpenCRISPR-1: both-strand NGG scanning, guide-format review, quality ranking and local specificity screening.
|
|
A machine-learning soft sensor that estimates penicillin concentration during fermentation, including batches where something goes wrong. Batch-wise validation, uncertainty and out-of-distribution detection. Read the preprint · Repository
|
A general SpCas9 and CRISPRi design workbench with NCBI and Ensembl sequence lookup, transparent guide ranking and local off-target screening.
|
More: AMS Venom Diagnostic Dashboard, an interactive simulation of antigen-scFv diagnostic signals · Basit Hub (code), my research start page with a 3D protein of the day and database search · all repositories
| Area | What I do in it |
|---|---|
| Phage and microbial bioinformatics | Endolysin discovery, phage protein annotation, host prediction, protein language models |
| Genome engineering | CRISPR guide design, multi-gene targeting, validation, sequence provenance |
| Bioprocess data science | Soft sensors, fault-inclusive modelling, uncertainty, out-of-distribution analysis |
| Structural bioinformatics | Protein structure analysis, molecular docking, molecular dynamics |
| Scientific software | Python tools, Streamlit apps, reproducible pipelines |
| Core | Python Git GitHub Linux Docker VS Code |
| Data and machine learning | NumPy Pandas SciPy scikit-learn Matplotlib Plotly Jupyter |
| Bioinformatics | Biopython HMMER NCBI Ensembl Streamlit |
| Structural biology | ChimeraX Schrödinger Glide HADDOCK AlphaFold / ESMFold |
| Web and testing | HTML CSS JavaScript Pytest GitHub Actions |
| Education | MSc Biotechnology (Industrial Biotechnology), Gujarat Biotechnology University |
| National examinations | GAT-B 2025, All India Rank 62 · CSIR-UGC NET, June 2026 · IIT JAM 2025, All India Rank 875 |
| Fellowship | GAT-B National Fellowship, Department of Biotechnology, Government of India |
| Preprint | Fault-inclusive soft sensing of penicillin concentration, bioRxiv, 2026 |
| Research experience | Structure-based study of human β-cardiac myosin inhibitors, IIT Roorkee |
| Recognition | Best Student of the Year (2022-24), MBSI with HiMedia Laboratories |
- Biologically grounded - assumptions and limitations are stated clearly
- Reproducible - versions, provenance, benchmark inputs and tests are kept
- Interpretable - outputs come with the context needed to judge them
- Usable - researchers can run a workflow through a practical interface and export the results
- Open to validation - a computational prediction is evidence to investigate, not a substitute for an experiment
I'm interested in research collaborations where biotechnology, computational biology and practical scientific software meet. If you have a biological question worth building on, write to me.
