Vaibhav Kulkarni

Vaibhav Kulkarni

Engineering Lead | Software, Data & ML Engineering for Computational Biology

PhD Computer Science · self-taught computational biologist & bioinformatician

📍 Switzerland

about

I am a software engineer by training, with a PhD in Computer Science & master's degrees in Embedded Systems & Information Systems. For over a decade I have built production-grade software, data platforms & machine-learning systems across life sciences, healthcare, supply chain & IoT.

My career spans academic research, peer-reviewed publishing & industry. That mix is what I bring to the table: the rigour of research, combined with the pragmatism of shipping systems that people rely on every day.

Along the way I turned these skills to biology & taught myself computational biology & bioinformatics. At Debiopharm I lead the data & bioinformatics team & work hands-on across preclinical trial analytics, biomarker discovery, antibody developability for our antibody-drug conjugate (ADC) programs & simulation-based design of bioequivalence trials.

I enjoy building cross-functional teams of software engineers, machine learning engineers, data scientists, bioinformaticians & computational biologists, where each discipline sharpens the others.

I believe in continuous learning: I contribute to open-source software & compete in computational biology challenges. At the heart of my work is a commitment to high-quality software, guided by a passion for clean, pragmatic code.

my path

  1. 01 · 2011–2015

    Embedded systems & IoT

    BOLT IoT · TU Berlin · TU Eindhoven · ETH Zürich

  2. 02 · 2015–2019

    Machine learning research

    PhD Computer Science · University of Lausanne

  3. 03 · 2019–2022

    Software & data engineering

    GenLots · SOPHiA GENETICS

  4. 04 · 2022–today

    Data, AI & computational biology

    Debiopharm

computational biology

Alongside building & running our data & analytics platforms, from the central data repository to our data warehouses, these are the scientific problems my team & I work on every day, from antibody design to the clinic.

Illustrations use simulated example data.

CH2CH3CH2CH3VHCH1VLCLVHCH1VLCLADC · DAR 4sequence liabilitiesEVQLVESGGGLVQPGGSLRLSCAASGFNGSYAMSN-G deamidation riskpredicted CMC riskHICTmPRAC-SINStiter

protein engineering

Antibody developability for ADCs

In-silico developability & design support for antibody-drug conjugate programs: protein language models, structure modelling & PTM-liability scans to rank CMC risk across antibody candidates & engineered formats.

10⁻³10⁻²10⁻¹11010²10³050100concentration (nM)viability (%)IC₅₀ADCnaked mAbisotype ctrl

in-vitro pharmacology

In-vitro assay analytics

Analysis of in-vitro assay data across our programs: cytotoxicity dose-response & IC50, antibody internalization & target binding, with standardized curve fitting & quality control across assays & CROs.

07142128050010001500days post-dosingTV (mm³)vehicletreated

in-vivo pharmacology

Preclinical trial analytics

Lead the data analysis & visualization for all preclinical efficacy trials at Debiopharm: harmonizing CRO data, computing tumor growth inhibition, AUC & relative tumor volume, & serving interactive dashboards across PDX & CDX mouse models.

-4-20240510log₂ fold change−log₁₀ p↑ up-regulated↓ down-regulated

translational oncology

Biomarker discovery

Transcriptomics biomarker discovery & validation for clinical-stage oncology programs: RNA-seq QC & batch-effect assessment, differential expression, pathway enrichment, ML gene signatures & omics foundation-model embeddings.

IHC tile 512×512normalize → segment→ scoreKi-67 positivenegativeKi-67 index

digital pathology

IHC image analysis

Pipeline for immunohistochemistry whole-slide images: tiling & stain normalization, CNN nuclear segmentation & Ki-67 scoring, benchmarked against a classical colour-deconvolution baseline.

071421280123time (days)conc. (ng/mL)observedsynthetic

clinical pharmacology

Synthetic data for bioequivalence

Generative models (VAE, WGAN-GP, physics-informed VAE, copulas) that learn from a small set of real PK profiles, with Monte Carlo trial simulations to test Type I error & optimize trial sample size.

FASTQQCalignquantvariantsTPMHRD·MSITMBDE·GSEAML sig.RNA-seq · variant calling · ctDNACRO & vendor data → central omics hub

bioinformatics

Genomics pipelines & omics hub

Built RNA-seq, variant-calling & ctDNA pipelines with HRD, MSI & TMB scoring, & automated transfers from CROs & sequencing vendors into a central omics hub alongside TCGA, CCLE & CPTAC.

competitions & open source

Antibody Developability Prediction

Ginkgo Datapoints AbDev Competition · 2025

Predicted 5 developability properties from sequence with ensembles of protein language model embeddings (ProtT5, p-IgGen, ESM-2) & LightGBM / ridge stacking.

Competition →

Adaptive Immune Profiling

Kaggle · AIRR-ML-25

Predicted immune state from T/B-cell receptor repertoires using k-mer XGBoost baselines & ESM-2 multiple-instance learning.

Competition →

Awesome Biopharma Foundation Models

Open source

Curated catalogue of foundation models for biopharma R&D: single-cell & genomics, antibody sequence & structure, generative design, small molecules & ADCs.

GitHub →

Mobility Prediction Literature

Open source

Curated research on human mobility modeling & next-place prediction, started during my PhD. My most-starred repository.

GitHub →

Geospatial Precipitation Pipeline

Open source

Transforms ERA5 precipitation data into queryable Parquet with geospatial indexing.

GitHub →

More on GitHub

Paper code & engineering notes

Code for my SIGSPATIAL & SGX papers, data-driven tech choices for a clinical data warehouse & a PostgreSQL tuning cheat sheet.

github.com/vaibhav90 →

experience

Debiopharm

Engineering Manager: Data & Bioinformatics

Debiopharm

July 2025 – Present
Lausanne, Switzerland
  • ▹Leading cross-functional team of data engineers & bioinformaticians to architect & deliver next-generation data fabric & bioinformatics platforms
  • ▹Leading preclinical analytics across in-vitro assays (cytotoxicity, internalization, binding) & all in-vivo efficacy studies, & transcriptomics biomarker discovery for clinical-stage oncology programs
  • ▹Delivering in-silico antibody developability & CMC risk assessment for antibody-drug conjugate (ADC) programs, using protein language models & structure prediction
  • ▹Built & deployed a dozen+ R Shiny & Streamlit applications on Posit Connect for translational medicine: in-vivo efficacy, cross-study pharmacokinetics & public omics exploration
  • ▹Engineered a data warehouse & monitoring dashboard for an antibody discovery program, automating ingestion of CRO deliverables
  • ▹Built RNA-seq, variant-calling & ctDNA pipelines (HRD, MSI & TMB scoring) & automated secure multi-omics transfers from CROs & sequencing vendors
  • ▹Prototyped generative models & Monte Carlo simulations for bioequivalence trial design, & an IHC digital-pathology pipeline for Ki-67 scoring
Debiopharm

Data Engineering Team Lead

Debiopharm

July 2022 – June 2025
Lausanne, Switzerland
  • ▹Architected & led the full-stack development of the company's most utilized data platform; Central Data Repository, automating ingestion & quality control of clinical & non-clinical trial data at enterprise-scale & positioning it as central data hub for all research & development programs
  • ▹Engineered the deployment of custom-configured scientific analytics platforms (SAS Viya, Posit, Dotmatics, Expressions) on company-controlled cloud infrastructure, creating a unified ecosystem interfaced with the central data repository to ensure stringent performance & data governance
  • ▹Engineered an entity-recognition & search platform by mining scientific publication databases - Automated internal trend-monitoring functions
  • ▹Productionised privacy-aware large language model-based applications to streamline document analysis & study protocol writing processes
  • ▹Supported compliance team via DevOps-based computerised system validation, infra qualification & security audits - Automated GxP workflows
  • ▹Responsible for technical due-diligence of healthcare-startups (technical-stack, software & ML strategy) to aid our VC funds investment decisions
SOPHiA GENETICS

Senior Data Engineer

SOPHiA GENETICS

April 2021 – June 2022
Lausanne, Switzerland
  • ▹Implemented terabyte-scale ETL pipelines to transform raw genomics data in-to analytics-ready format - Significantly lowering processing time
  • ▹Developed API endpoints & CLI tools to expose database functionalities to internal teams- Facilitated data democratisation & access control
  • ▹Implemented regression testing mechanisms for variant detection bioinformatics pipelines & dashboards delivering real-time insights into KPIs
  • ▹Authored formal TLA+ specifications for ETL operations, enabling quantification & benchmarking of logical correctness across the data stack
GenLots

Software Engineer | Research & Development Lead

GenLots

September 2019 – March 2021
Lausanne, Switzerland
  • ▹Developed a microservices architecture, streamlining ingestion, transformation, & delivery of client ERP data to production-grade ML services
  • ▹Deployed reinforcement learning-based solution, optimising end-to-end supply chain planning, warehouse management, & CO2 emissions
  • ▹Engineered a suite of tools for reporting & monitoring integrated into dashboards delivering real-time business performance metrics to clients
  • ▹Directed research projects in collaboration with Swiss watch manufacturing firms to predict inbound material needs & stabilise supply chains
HEC Lausanne

Research Scientist

Distributed Object Programming Lab

November 2015 – August 2019
Lausanne, Switzerland
  • ▹Partnered with public transport firms to architect & implement systems capturing high-frequency, real-time ridership data into a data warehouse
  • ▹Developed a Machine Learning-driven platform to proactively detect & reduce ticketless travel across the Lausanne public transport network
  • ▹Led platform development for a large-scale (>300 participants) spatiotemporal mobility-data collection project to facilitate ML-privacy research
ETH Zürich

Project Engineer

ETH Zürich

January 2015 – August 2015
Zürich, Switzerland
  • ▹Implemented a software-defined-radio-based platform for reliable IoT product testing, generating interference patterns of wireless appliances
  • ▹Deployed the platform in production at ETH Zurich & TU Berlin with toolsets allowing for remote configuration & logging performance metrics
BOLT IoT

Embedded Systems Engineer

BOLT IoT

September 2011 – January 2013
Goa, India
  • ▹Designed PCB layouts for embedded platforms used in robotics & IoT systems & collaborated with software teams to ensure firmware integration
  • ▹Implemented toolsets & applications to benchmark platform performance, validation tests, EMI simulations & industry compliance certifications
NIO

Project Intern

National Institute of Oceanography

June 2011 – August 2011
Goa, India
  • ▹Contributed to autonomous underwater vehicle localization project by developing triangulation algorithms & learning satellite communication & ocean current modeling

skills

▶ Software & Data Engineering

Software Engineering

Writing clean, tested, maintainable software across the stack, from embedded firmware to web applications & data services.

Python R TypeScript SvelteKit C# / .NET Rust Java Scala SQL C (embedded) Test-Driven Development REST APIs

Enterprise Platform Architecture

Designing & building scalable, production-grade platforms from data ingestion to user-facing applications, with focus on system reliability, performance optimization, & maintainability.

Distributed Systems Microservices Architecture Domain-Driven Design (DDD) Onion Architecture API Design System Integration

Data Architecture

Designing & implementing enterprise-scale data repositories, data warehouses, & real-time data processing systems.

Data Warehousing Data Lakes SQL (PostgreSQL) NoSQL (MongoDB) ETL/ELT Pipelines DuckDB dbt Apache Airflow Star-Schema Modeling

Cloud & Infrastructure

Deploying & managing scalable, cloud-native applications & infrastructure.

AWS Azure GCP Docker Kubernetes Terraform Infrastructure as Code (IaC) Terragrunt S3 ↔ Azure Blob Transfer Microsoft Graph API
▶ AI & Machine Learning

Production ML Systems (MLOps)

Building & deploying robust, scalable, & reproducible ML pipelines & services.

PyTorch Scikit-learn MLflow Kubeflow CI/CD for ML

Large Language Models (LLMs)

Developing & productionizing privacy-aware LLM applications for scientific document analysis & generation.

Hugging Face LangChain Vector Databases Prompt Engineering

Protein & Omics Foundation Models

Applying protein language models, structure predictors & omics foundation models to antibody engineering & biomarker discovery.

ESM-2 ProtT5 AbLang-2 p-IgGen SaProt IgFold AlphaFold Geneformer BulkFormer

Predictive & Generative Modeling

Creating ML-driven platforms for predictive analytics & statistical analysis.

Python (Pandas, NumPy) Jupyter Reinforcement Learning XGBoost LightGBM Multiple-Instance Learning VAE / WGAN-GP Physics-Informed Neural Networks Computer Vision (U-Net)
▶ Computational Biology & Bioinformatics

Multi-Omics Data Processing

Engineering bioinformatics pipelines for harmonizing & analyzing genomics, transcriptomics, & other omics data.

Nextflow Snakemake Bioconductor RNA-seq & Variant Calling Pipelines STAR MultiQC PyDESeq2 GSEApy / MSigDB ComBat HRD / MSI / TMB ctDNA / Liquid Biopsy TCGA · CCLE · CPTAC

Antibody Engineering & Developability

Sequence- & structure-based assessment of antibody developability for ADC programs.

ANARCI / IMGT CamSol TANGO Foldseek PTM Liabilities Antibody Format Engineering

Preclinical Pharmacology & PK

In-vitro assay & preclinical trial analytics, pharmacokinetic studies & simulation for clinical trial design.

Dose-Response & IC50 Internalization & Binding Assays PDX / CDX Efficacy Studies TGI · AUC · RTV PK / NCA PopPK Simulation Bioequivalence Statistics

Cheminformatics

Supporting small molecule & linker-payload research with data analysis & molecular modeling.

RDKit Cheminformatics Molecular Modeling Tools

Biomarker Discovery

Developing platforms & strategies to identify & validate novel biomarkers from complex datasets.

R Python Statistical Analysis Data Visualization Digital Pathology (WSI, IHC)
▶ Quality & DevOps

Automated Testing Frameworks

Developing & owning regression & performance testing systems for critical data pipelines.

Pytest CI/CD Integration GitHub Actions Jenkins

System Validation & Qualification

Implementing DevOps-based automation for GxP compliance, infrastructure qualification, & security audits.

GxP Compliance DevOps Methodologies Automated Auditing

Formal Methods

Using formal specifications to verify the logical correctness & reliability of complex data systems.

TLA+
▶ Leadership & Strategy

Cross-Functional Team Leadership

Leading teams of data engineers, bioinformaticians, & software developers.

Agile Scrum Kanban Jira Confluence

Technical Strategy & Roadmapping

Defining the technical vision for AI & data platforms in R&D environments.

Product Roadmapping Stakeholder Management Vendor & Platform Selection

Technical Due Diligence

Evaluating the technology stacks & ML strategies of healthcare startups for investment purposes.

Tech Stack Evaluation Risk Assessment

education

UNIL

PhD in Computer Science

University of Lausanne - Information Systems Department

November 2015 – August 2019

Thesis: Information Systems for Privacy-Aware Machine Learning

ETH Zürich TU/e

M.Sc. Embedded Systems

ETH Zürich & Eindhoven University of Technology

August 2014 – August 2015

Master thesis at ETH Zürich · Graduated cum laude

TU Berlin

M.Sc. Information & Communication Technology

Technische Universität Berlin

February 2013 – August 2014

Graduated with Honours

peer-reviewed publications

Generating Synthetic Mobility Traffic using Recurrent Neural Networks

V. Kulkarni, B. Garbinato

ACM SIGSPATIAL - AI and Deep Learning for Geographic Knowledge Discovery, 2017

PDF →

Examining the Limits of Predictability of Human Mobility

V. Kulkarni, A. Mahalunkar, B. Garbinato, J. D. Kelleher

Entropy, 2019

PDF →

Generative Models for Simulating Mobility Trajectories

V. Kulkarni, N. Tagasovska, T. Vatter, B. Garbinato

Neural Information Processing Systems Workshop on spatiotemporal modeling, 2018

Preprint →

MobiDict: A Mobility Prediction System Leveraging Realtime Location Data Streams

V. Kulkarni*, A. Moro*, B. Garbinato (*co-primary authors)

ACM SIGSPATIAL Workshop on GeoStreaming, 2016

PDF →

Information Disclosure in Location-based Services: An Extended Privacy Calculus Model

D. Naous, V. Kulkarni, C. Legner, B. Garbinato

International Conference on Information Systems (ICIS), 2019

PDF →

Privacy-Preserving Location-Based Services by using Intel SGX

V. Kulkarni, B. Chapuis, B. Garbinato

ACM SenSys Workshop on Human-centered Sensing, Networking, and Systems, 2017

PDF →

20 Years of Mobility Modeling & Prediction: Trends, Shortcomings & Perspectives

V. Kulkarni, B. Garbinato

Advances in Geographic Information Systems (ACM SIGSPATIAL), 2019

PDF →

Capturing Complex Behavior for Predicting Distant Future Trajectories

B. Chapuis, A. Moro, V. Kulkarni, B. Garbinato

ACM SIGSPATIAL Workshop on Mobile Geographic Information Systems, 2016

PDF →

On the Inability of Markov Models to Capture Criticality in Human Mobility

V. Kulkarni, A. Mahalunkar, B. Garbinato, J. D. Kelleher

28th International Conference on Artificial Neural Networks (ICANN), 2019

PDF →

Extracting Hotspots without A-priori by Enabling Signal Processing over Geospatial Data

V. Kulkarni, A. Moro, B. Chapuis, B. Garbinato

Advances in Geographic Information Systems (ACM SIGSPATIAL), 2017

Preprint →

Breadcrumbs: A Feature Rich Mobility Dataset with Point of Interest Annotation

A. Moro, V. Kulkarni, P. Ghiringhelli, B. Chapuis, K. Huguenin, B. Garbinato

Advances in Geographic Information Systems (ACM SIGSPATIAL), 2019

PDF →

Controlled Interference Generation for Wireless Coexistence Research

A. Hithnawi, V. Kulkarni, S. Li, H. Shafagh

ACM MobiCom workshop in Software Radio Implementation Forum (SRIF), 2016

PDF →

Capstone: Mobility Modeling on Smartphones to Achieve Privacy by Design

V. Kulkarni, A. Moro, B. Chapuis, B. Garbinato

IEEE Conference On Trust, Security & Privacy In Computing & Communications, 2018

Preprint →

A Mobility Prediction System Leveraging Realtime Location Data Streams

V. Kulkarni, A. Moro, B. Garbinato

ACM Conference on Mobile Computing and Networking (MobiCom) (Poster), 2016

PDF →

certifications

Agentic AI and AI Agents

Vanderbilt University

January 2026

View Certification →

Finetuning Large Language Models

DeepLearning.AI

January 2026

View Certification →

Machine Learning in Production

DeepLearning.AI

July 2025

View Certification →

Drug Discovery

UC San Diego

September 2024

View Certification →

Rust Fundamentals

Duke University

June 2024

View Certification →

Generative AI with Large Language Models

DeepLearning.AI & AWS

December 2023

View Certification →

Data Engineering with Azure

Udemy

January 2023

View Certification →

Applied Machine Learning

University of Michigan

April 2020

View Certification →

Data Science

University of Michigan

April 2020

View Certification →

Amazon Web Services

AWS

May 2019

View Certification →

Laws & Economics of Media Platforms

University of Chicago

June 2018

View Certification →

Information Security

University College London

January 2018

View Certification →

Machine Learning

Stanford University

January 2016

View Certification →