The Berkeley Science Fellows Program: On campus professional development opportunity for postdocs



The mission of the program is two-pronged‐ provide career transition opportunities to postdocs by helping them experience working with startups; 2. Helping startups solve scientific problems and reach their milestones while being incubated on campus.

Our team

Out team includes: Naresh Sunkara Pooja Mukherjee.


The core function of the Berkeley Science Fellows Program is to connect postdoctoral researchers with startups in the UC Berkeley entrepreneurial community. This will provide valuable real world working experience to postdocs while helping the startups benefit from the scientific expertise of the postdocs.

Berkeley Science Fellows will offer their advice to startups to solve specific scientific challenges including product development, process trouble shooting, grant writing/reviews, experimental data analysis, and presentations, to participate in discussions in brainstorming sessions to develop start up ideas.

The Berkeley Science Fellows Program will be responsible for recruiting postdocs (after an interview) and present them to startups at SkyDeck. Science Fellows will be given an opportunity to select the startup they would like to work with as well.


  • Candidates must be a postdoc at UC Berkeley to be eligible to participate.

Time Commitment

It is understood that postdocs have a primary commitment to their research at UC Berkeley.  Science Fellows will need to commit 4-5 hours a week to assist the startup. The hours are flexible and may include evenings and weekends. Please note that the Science Fellows are not compensated for their volunteer assistance.

Because postdocs will gain valuable experiences while working in a fast paced startup environment, the selected postdocs are expected to deliver on their agreed-upon time commitments. To this regard, the Science Fellows will be intermittently contacted by the Program leads to assure that all commitments are being met.  Additionally, there will be monthly participant check-ins to solicit feedback to improve the Berkeley Science Fellows Program. 

Benefits to postdocs

  • Gain experience working with startups
  • Get exposure to potential employment opportunities
  • Add work experience to your resume
  • Receive a letter of recommendation from a startup
  • Serve on the Science Fellows Advisory Board (after completion of internship)
  • Receive satisfaction of helping a startup succeed

Benefits to startups

  • Access to world class talent (free of cost)
  • Future Employee pipeline
  • Help with troubleshooting specific problems
  • Help with grants (review/write)


  • This is an unpaid and volunteer program

Important links

Link for applicants to register as a Science Fellow:

List of startups 

  • RJH Biosciences Inc.
    • We are solving the clinical delivery problem of nucleic acid therapeutics.
    • Skills Needed:
      • PhD in chemical engineering, chemistry, pharmaceutical sciences or related field.
    • Specific Project:
      • We want to develop a technology for COVID-19 therapy based on siRNA.
    • Resource:
      • See Uludag et al., Prospects for RNAi Therapy of COVID-19. Front Bioeng Biotechnol. 2020 Jul 30;8:916. doi: 10.3389/fbioe.2020.00916.
  • Kintsugi Mindful Wellness
    • Kintsugi is creating a novel voice-based AI technology to measure and predict well-being. Currently we can detect depression and anxiety from less than a minute of free-form speech with twice the accuracy of a general practitioner.
    • Skills Needed:
      • Model performance optimization, error analysis, transfer leaning, data augmentation, speech processing, representation learning, various deep learning architectures
    • Specific Project:
      • Model performance optimization, error analysis, transfer leaning, data augmentation, speech processing, representation learning, various deep learning architectures
  • Jubily
    • Finding culturally competent clinicians in mental health and medicine can be such a challenging, daunting, and cumbersome process that often relies on word of mouth referrals, multiple phone calls and long, page-by-page searches on directories. Suffering with an illness can be a lonely experience. For clinicians, finding outside referrals can also be a cumbersome task that's often limited to one's personal professional network.
    • We simplify the process of finding culturally competent healthcare providers, while providing online communities of peer support to help you thrive by leveraging technology using our research-based algorithm
    • Skills Needed:
      • I'm looking for someone with experience building consumer facing mobile applications for iOS and Android. 
    • Specific Project:
      • Backend app development (iOS and/or android). 
  • Nimble
    • Real time analytics of e-sports making predictions for in-game events.
    • Skills Needed:
      • Computer vision and understanding on numpy
      • Machine learning and statistical analysis
    • Specific projects:
      • Computer vision for extracting real time information from game footage: E-sports generate a multitude of data through their video feed and using computer vision techniques (optical flow to OCR) allowed extraction of these important features that can be used for making in game predictions.
      • Modelling e-sports: Unlike traditional sports e-sports can not be modelled by just providing more data to a deep learning network. The innumerable rule changes and new additions makes it impossible to do so, but rather a new paradigm of modelling is required.
  • Glucosia, Inc.
    • Producing a commercial product
    • Skills Needed:
      • Biosensor Data Acquisition Development, C++, C#, App Development, Electrical Engineering
    • Specific projects:
      • Biosensor Development, Hardware Implementation, Grant Writing
      • Job Duties & Responsibilities: • Understand biosensor specifications and capabilities • Design and implement software of embedded devices and systems from requirements to production and commercial deployment • Architect, design, code, and document firmware for electronic and electromechanical products • Code accurate and efficient machine learning algorithms minimizing computational power usage • Bare-metal as well as RTOS-supported development experience in C/C++, ARM processors a plus • Experience developing real time electromechanical control systems and signal processing • Write software requirements, design documents, and test procedures • Work closely with the rest of the engineering team to integrate designs and solve problems. • Provide timelines with milestones and concrete deliverables • Participate in technical design reviews • Support software QA and optimize I/O performance • Design on PCB design and then following through to manufacture, utilization & characterization, next-generation equipment & devices • Performs assignments of a very complicated or difficult nature requiring a high degree of technical competence. • Build prototypes both independently and under the guidance of co-workers. • Research clinical literature to ensure that customer requirements are met. • Work closely with Q/A to develop standards, perform verification and validation testing, perform failure analysis, and to continuously improve the company’s products and processes. • Work effectively within the confines of a regulated, quality system driven, FDA monitored medical device environment. • Create manufacturing methods that are developed, characterized, bounded and repeatable. • Demonstrated desire to work in a small medical device startup. • An inventive aptitude balanced with discipline and detail-orientation. • A willingness to work hard, with a driven outlook and with a strong desire to succeed as an engineer, as a team, and as a company. • Open to team building exercises around food, especially food trucks
  • Pyoneer
    • Extracting the insights out of a mass of feedback data
    • Skills Needed:
      • Vectors, Vector Calculation, Databases (Postgres), kNN Algorithm
    • Specific problems:
      • Similarity Calculation of Vectors (in a Software Product)
  • FixingCO2
    • Building a reactor that converts CO2 into fuels and chemicals
    • Skills Needed
      • Electrochemistry experience
    • Specific problems:
      • Scaling up CO2 electrolyzer operations by modifying PEM hydrogen electrolyzers. Testing and optimizing the reactor
  • Scopi
    • Nasoendoscopes are tiny cameras that are used to diagnose speech disorders. They cost upwards of $80K. Because of this high cost, 80% of speech clinics in developing countries do not have access to a nasoendoscope and cannot diagnose their patients. Their patients cannot get any treatment and have long-lasting speech problems. We created a design for a nasoendoscope that costs 5% of what current systems cost and improves upon existing nasoendoscope design by being portable and comfortable (for patient and physician). We are a team of engineers, economists, and physicians whose mission is to provide diagnostic devices to every speech and sleep physician around the world, especially those located in low-resource areas. We work closely with and are supported by Stanford University's Medical and Engineering schools, medical device accelerator ZeroTo510K, UC Berkeley's startup incubator SkyDeck, and international business competition Fishbowl Challenge. We partner with international cleft NGO SmileTrain.
    • Potential Projects:
      • EE - Image Processing Pipeline: We currently use a very basic image processing pipeline (IPP). We'd like this person to review possible IPP options (eg white balancing or noise algorithms), suggest which need to be included, and (if time permits) implement them. Also includes demosaicing. Must have experience with imaging, C/C++ coding, and ideally with STM32 IDE.
      • Med/BioTech: Writing Test Procedures (and if time permits analyzing results) for non-Inferiority cadaver and anatomical model tests (tests to be done at Stanford Medical Center) as well as a durability "shake rattle roll" test. This is to test our v2 prototype. Ideally, this person has a medical (or biomed) background, ideally also regulatory or quality control/assurance expirience.
      • EE - PCB design review: Our design has 3 custom PCBs which are some of the most expensive components in the BOM. This person would analyze and optimize the PCB design (eg change the way we depanel the boards, move around vias, replace components, change vendors, etc). Must have a PCB background, ideally with HDI/micro PCB and laser depanelization experience. Our PCBs are very small.
      • EE - Optimize STM32 processor coding: The code for our STM32 processor could be optimized to get a faster video speed. This person would identify what can be improved and if time permits write this code. Must have experience with C/C++ coding, and ideally with STM32 IDE and image sensors/cameras.

Serinus Labs, Inc.

We are developing early warning sensors to address the problem of lithium-ion battery safety in electric vehicles.

Skills Needed: MEMS gas sensors, silicon CMOS or MEMS processing, hardware development (PCB, ASIC), Reliability modeling, FEM modeling and simulation.

Specific Projects: Our current projects of interest. Interest & expertise in any of these areas will be useful:

1) Development of high throughput scheme for characterization of chemical sensitive field effect transistor gas sensors to develop a fast calibration model.

2) Reliability modeling of chemical sensitive field effect transistor gas sensors.

3) Electro-thermal FEM modeling for sensor chip layout-power optimization.

4) Optimization of Serinus Gen-1 COTS (Arduino) evaluation hardware.

5) Analog front-end ASIC interface development (this will primarily involve just interface with ASIC company to understand capabilities and come up with a BOM and cost per chip estimate).

6) Silicon CMOS and MEMS processing foundry interface. (This will primarily involve interface with an foundry to gauge process compatibility).  

7) Battery abuse test bed development (essentially developing a list of tools and identifying potential locations/labs for a future company lab).

NuPort Robotics Inc.

NuPort Robotics is addressing the problem of severe driver shortage and increasing costs of drivers in the trucking industry. We are building technology for self driving trucks to automate the middle mile transportation. Our mission is to automate global supply chains using self-driving trucks increasing sustainability through the adoption of clean transportation solutions and reduce the cost of operations for clients in retail, logistics, and automotive manufacturing. 

Skills Needed: 

-  Strong background in Autonomous Driving, Robotics, Computer Vision, Machine Learning, Deep Learning

-  Experience in ROS, C++ and / or Python

-  Proficiency in Linux

-  Experience working with various sensor modalities: Lidars, Cameras, Radars, GPS

-  Experience in a subset of the following topics: 3D mapping, localization, perception (object detection, localization, tracking), prediction, motion planning, motion estimation, controls, sensor fusion, mapping sensor calibration, etc.

Specific Projects: 

1. Motion Planning and Controls for Autonomous Driving: a basic overview of the role can be found here: 

We have a Senior Controls Engineer who has a PhD in Autonmous Driving, we are looking for someone who can complement our engineer with prior experience in Motion Planning.

2. 3D Mapping: NuPort Robotics is looking for assistance in creating high definition maps based on multiple camera data, 3D Lidar Data, and continually updating the maps based on detection of changes in the map. 

More details are available here: 

3. Prediction Role (perception focus): Assistance is required on the following topics

- build new deep learning models that can predict the future behavior of other road users in real-time.

- develop efficient traditional and deep learning architectures that can run in real time

4. Perception Role: Assistance and support is required on the following topics:

Sensor Fusion based approaches for Object detection, localization, tracking, road segmentation, etc. 


Videos are exploding. By 2021, over 1M minutes of videos are said to cross global IP servers every second. Given this high volume and the increasing trend of livestreams, it is becoming increasingly difficult to tag, index and search through these videos at scale. 

We are developing next generation video search. What does that mean? Searching for a specific footage inside large video libraries by using text keywords. For this purpose, we are developing a vision AI engine, which is capable of producing the most relevant metadata and giving the context in video. We are combining different modalities to extract the most contextual video understanding. 

Skills Needed:

Project 1: 

- Having great understanding of computer vision and machine learning approaches on video understanding

- Experience on real time computer vision. 

- Experience in theoretical and empirical research and for addressing research problems

- Knowledge in a programming language

- Contribute to computer vision systems deployed on commerce projects across Verchabe

- Contribute to computer vision systems deployed at Verchable

- Plan and execute cutting-edge research to advance video understanding

Project 2:

- Having a great understanding of action recognition, video tagging, video understanding as sequence modelling

- Knowledge in deep learning and video understanding methods 

- Knowledge in a programming language

- Contribute research that can be applied to Verchable product

- Contribute to computer vision systems deployed at Verchable

- Experience in theoretical and empirical research and for addressing research problems

- Plan and execute cutting-edge research to advance video understanding

Answer Job3:

- Having a great understanding of information retrieval, ranking and NLP

- Plan and execute cutting edge research to advance ranking system 

- Knowledge in programming languages and databases

- Contribute to the ranking and retrieval system at Verchable 

- Communicating and collaborating with CV&ML researcher 

- Optimizing DNN, hands on experience on video understanding 

- Experience on real time computer vision

Specific Projects:

Project 1

Real time video understanding: Real-time vision: creating an optimized computer vision AI, optimizing models and overall algorithms in the pipeline (Action Recognition and Multiple Object Tracking). This system will operate in real time and the main purpose is collecting relevant metadata in real time from smart home cameras. In order to get low latency, we are considering deploying our vision AI on edge by using sophisticated chipsets with combination of cloud computing. 

Project 2

Advance Video understanding: Generalistic Action recognition, video tagging and combination of cutting edge CV for large video libraries. The aim is extracting the contextual metadata from video libraries and making them searchable. 

Project 3

Creating an Information retrieval, search system and ranking which is essential for our search engine. In this part of the project we are aiming to create a database which is searchable by using keywords and the corresponding results are video footages

Frontier Microscopy

Technicians in laboratories across industry quietly make the world go around. Every day tens of thousands of lab techs are pushed to their limits conducting repetitive, but critical, manual microscopy analyses for all major industries such as pathology, environmental services, agriculture, and mining. Results are often required as quickly as possible, which often means pushing through the night to get things done.

Our vision is for a future where scientists are given real opportunity to do their best work whilst also keeping industry moving. That's why we're building MARVIN, an automated microscopy platform that will allow the worlds scientists to automate essential microscope analyses. MARVIN combines our own robotic microscope and configurable AI software under a subscription model that is designed for accessibility.

Skills Needed: Computer Vision, Machine Learning, Robotics.

Specific Projects: The successful applicant would most likely work towards applying MARVIN for a new application in automated microscopy such as for water quality or pathology. Ideally they should have some experience performing work in a laboratory setting but otherwise must be highly competent in computer vision and machine learning technologies.

They may be proficient in any programming language and will not be required to deliver anything more than a proof-of-concept or prototype.

Hopeful Inc. is a company that helps nonprofits generate donations from new people by telling compelling stories. We do this through a machine learning driven analytics dashboard that enables nonprofits to analyze trends in their social media strategy, monitor their supporters’ engagement, and proactively create programs to better drive donations  through data-driven decision making.

Skills Needed: Computer Science,Data Science,Statistics,Machine Learning,Artificial Intelligence,Python,R,Tensorflow,Keras,Scikit-Learn,Numpy,Pandas,Google Cloud Platform


Hopeful Inc. is seeking to develop its Storytelling Artificial Intelligence (SAI) feature, a subset of the

greater Hopeful Inc. Data Analytics Platform. SAI allows nonprofit leaders, board members, and nonprofit marketers to aggregate the content they create for their nonprofit marketing strategies – specifically case studies and stories they tell of the impact they have. The goal of Storytelling AI is to ultimately allow nonprofit communications teams to automatically create content based on the quantitative output of our machine-learning driven algorithm, so that there content is immediately optimized for donating donations.

This aggregation encompasses:


Direct Posts

Instagram Stories

Instagram TV (Long-form posts)

Sponsored Ads


Direct Posts

Facebook Stories

Facebook Events (?)

Google SEO – Website

Blog Posts

Banner Ads


RSS Feed










SAI stories can be tracked by users in single components, allowing the user to see which stories are performing better than others across all major social media platforms. Furthermore, these stories are categorically ranked by their impressions, reach, and average amounts of likes/comments, allowing the user to see which stories are creating the most significant ROI.

Additionally, SAI will have actionable insights provided by a proprietary algorithm built in-house by the Hopeful technology team. This algorithm will dynamically learn what stories are best fit for certain audiences, events, and platforms, and provide in-app suggestions to the user on how their stories can be best told

Customer Workflow

Customer logs into Hopeful Platform using username/password

Customer navigates to the ‘Account Setting"

Polybiomics is changing analysis of live cells – the simplest proxies for humans and the forefront of multi-billion dollar pharma and clinical research markets. Currently, researchers take sequential measurements of separate cell samples at variable timepoints using multiple instruments and then try to correlate unlinked data. These disparate measurements and data are hard to assemble into a comprehensive view of cell health and function. This affects the quality of therapeutic development, especially in immuno-oncology, where live cells are used as medicine. 


Scientific reviews and customer interviews have emphasized on the pressing need for maximizing knowledge gain from current cellular analyses by running information-rich cellular assays to derive actionable insights that predict behavior of therapies, more accurately.


Skills Needed: Cell biologist with understanding of biochemistry, Immunologist , Electrochemists

Optical engineer or physicist for biodevice optimization, Software/Data Scientist, Marketing in life sciences and biotech


Specific Projects:


We have R&D prototyping, early phase immunotherapy science and marketing projects that include the following areas:

1) Performing scientific and market research in:

o Cellular therapeutics, in particular: Immunotherapy manufacturing and endpoints/modalities that are used for assessing cellular therapeutics potency and efficacy and toxicity

o Biosensors for immune-oncology


2) Setting up benchmarking cell culture assays for:

o count, viability, morphological, and biochemical analyses of live cells  

o metabolism analysis of T-cells in immuno-oncology

o T cell killing assays that might involve co-culture

3) Engineering for biodevice design and fabrication 

4) Bio-signal crosstalk assessing

5) Software development for big data and AI