The Context the Interface Cannot Lose

SuperMe

A quiet operator can be useful to thousands of people and still dread writing a LinkedIn post. SuperMe is building a professional network around that person, with AI profiles that let others question the experience behind the work.

Casey Winters, co-founder and CEO, and co-founder Ludo Antonov announced a $20M Series A on October 29. Kirsten Green at Forerunner led the round, with Greylock Partners participating. The investment backs a way to make professional knowledge travel without requiring its owner to take every call.

That participation problem deserves attention. Someone can understand a pricing decision, a growth experiment or a hiring mistake and have little interest in turning the lesson into a content strategy. A polished public presence takes another kind of work. Plenty of capable people already have a job.

SuperMe lets contributors build conversational profiles from material they choose to share. Users can bring several profiles into a group chat and examine different perspectives. The humans behind those profiles do not need to attend, and the output should be understood as AI-mediated guidance drawn from their experience.

The founders have spent years close to network behavior. Casey Winters worked across Grubhub, Pinterest and Eventbrite. Ludo Antonov led engineering at Pinterest, Lyft and Whatnot. Their new assignment puts the usefulness of an operator's knowledge inside the professional network itself.

There is a commercial bargain here: contribute something worth learning from, get discovered for it. For a builder asking a difficult question, the attraction is finding relevant experience beyond the people willing and available to take a meeting. For the contributor, recognition can begin before another introduction lands in the inbox.

SuperMe now brings those perspectives into Claude and ChatGPT. That makes distribution unusually practical: the question can stay alongside the context somebody has already assembled, while additional experience enters the discussion. The company says the funding will expand its expert network, improve matching and make access easier within existing tools.

Matching carries a lot of responsibility in that plan. Advice earns its place when the experience behind it fits the problem in front of someone. As SuperMe brings more operators into the network, Casey Winters and Ludo Antonov have to keep making that fit useful enough for people to share the work that shaped their judgment.

Creandum VIII

Fredrik Cassel once tried to build a Nordic version of Google's advertising business. In his Creandum biography, he describes a product that fell short, customers who weren't ready, and the disadvantage of building far from the best-funded competitors.

That experience now sits inside Creandum's partnership as the firm closes Creandum VIII, a $750M fund backing technology companies from pre-seed to Series A. The firm is extending its reach across Europe and the US while keeping its generalist approach.

Cassel joined founder Staffan Helgesson as an analyst when Creandum began in 2003. Today, the Stockholm-based Partner works within a firm with hubs in Stockholm, London, Berlin and San Francisco. A founder's location can put them near exceptional engineers while leaving them a long way from the customers, capital and introductions they need.

The new vehicle follows Creandum VII, which closed at €500M in June 2024. EU-Startups reports that VIII is expected to support roughly 30 investments over the next two years, with commitments from around 30 institutions in Europe and the US. Individual LPs and check sizes remain undisclosed.

The larger fund still leaves the partners deciding where conviction belongs. An early-stage company still arrives with an unfinished product, a small team and assumptions that customers have yet to settle. More places to find those companies also means more contexts to understand. The time a partner spends learning a market becomes part of the founder's financing decision, long after the amount has been agreed.

Peter Specht, the firm's London-based Partner, has written about that work through Creandum's investment in Rillet. His 2024 memo described the opportunity in rebuilding the general ledger for accounting automation. The useful detail was the job a finance team needed done, and the difficulty of replacing software it already depended on.

Creandum is applying the same demand for reinvention to itself. The partnership says it is rebuilding how it sources, decides and supports companies with AI at the core. The announcement offers no measured operating gains, but it does put the investor's own methods inside the change it expects founders to navigate.

For the next founder meeting Cassel or Specht, the conversation can travel from a local market to a global ambition very quickly. Somebody on the other side of the table still has to understand why this product, for these customers, deserves the firm's attention.

CareLumi

A hospital can keep its software and still become a CareLumi customer. The company's API route makes verification a purchase of its own, alongside its platform for getting clinicians ready to practice.

That is a useful commercial choice in healthcare. An organization may already have tools for credentialing, licensing, privileging and payer enrollment. The team responsible for the result still needs evidence that the work inside those tools is correct.

Shai Basys, CEO, Claire Chen, COO, and Eshan Saran, CTO, are building CareLumi around both assignments. The Chicago company announced an oversubscribed $2.5M pre-seed round on October 28, led by Izou Partners.

Fifth Star Funds, Longjump Ventures, Wellforge, SCS Innovations, University of Chicago Booth, Illinois DCEO and Chicago Early Growth Ventures also participated, according to CareLumi. The local relationships have some history: Polsky selected CareLumi for its 2025 Build accelerator and listed a $135K New Venture Challenge investment in June 2026.

The product gives the conversation somewhere concrete to go. CareLumi describes checks against primary sources, an action history and reports tracing a failure to the step that needs correction. A named reviewer approves a correction before it is recorded as resolved. That task includes the evidence needed to question the answer.

That reviewer has to work with what the software leaves behind. An alert without the original evidence simply hands another investigation to an already busy person. A usable source record can make the handoff more productive, especially when the question is whether a provider is ready to practice or bill.

CareLumi can sell the complete provider-operations workflow, or it can connect to AI and software a customer already uses. The second route gives the company a way into an organization without asking every existing application to move out first.

For Shai Basys, Claire Chen and Eshan Saran, that broadens the customer conversation considerably. A practice may want administrative work taken off its hands. A health system may want to understand and review work happening across tools it intends to keep. Each needs a clear account of who checks the output and who owns the correction.

The new capital supports provider readiness and healthcare AI verification. The work now travels into those organizations, where an approval still needs a person willing to put a name beside it.

Setara Biosciences

David Goldstein has a job to finish at Jazz Pharmaceuticals while helping build Setara Biosciences. After Actio's acquisition, its co-founder is staying with Jazz at least through acceptance of the ABS-1230 new drug application submission, while serving as Setara's president and executive chair.

That is an unusually concrete picture of what survives a biotech acquisition: scientific responsibility, spread across two companies with different programs to advance. Setara announced a $90M Series A on October 30 to carry forward Actio's remaining pipeline, with its dedicated capital and leadership.

Deerfield Management, Canaan, Droia Ventures, EcoR1, Euclidean Capital and Regeneron Ventures participated. Jazz became a shareholder through the Actio transaction. Investors familiar with the original work are financing another stretch of it, while the buyer keeps an interest in the science moving outside its walls.

Co-founder and CSO David Breckenridge joins Goldstein, alongside CBO Pete Condon and CFO Cayce Denton. The team is bringing forward STAR-0871, formerly ABS-0871, an investigational oral inhibitor of TRPV4, a protein channel whose excessive activity can cause Charcot-Marie-Tooth disease type 2C.

Setara says the healthy-volunteer portion of Phase 1 is complete. It plans the THERON Phase 1b patient study in 2027 and a separate Phase 2a study in overactive bladder. The same molecule now has two clinical assignments, reaching from a rare genetic disorder into a much more prevalent condition.

A clear genetic cause gives researchers somewhere to aim. Measuring whether treatment changes a slowly progressing disease is another demanding piece of the work. Goldstein described that difficulty to BioPharma Dive: several muscle groups can weaken, leaving investigators to determine how best to assess a patient's function and any improvement.

The financing supports that journey into patients, alongside earlier discovery programs in genetic epilepsies and migraine. Setara's wider ambition is to use insights from rare diseases to pursue related biology in larger populations. Each expansion still needs its own clinical evidence.

There is institutional value in keeping the people who understand a program close to its next decisions. Setara starts with that continuity, while Goldstein's work carries him between the epilepsy program advancing at Jazz and the patient studies being prepared under a new company name.

Wrapped Health

Max Russell-Harris had recovery instructions for his father. Finding the products those instructions called for became another family job, when the family already had enough to manage.

That experience became Wrapped Health, the Washington, DC startup connecting a clinician's recommendation with the purchase that follows it. On October 29, the company announced $2.1M in Pre-Seed funding co-led by Lightbank and Arteria Capital.

Good advice can still leave a family running procurement. Which product? Which size? Which account can pay for it? A family trying to help somebody recover can spend its energy resolving details the care team already understood.

Co-founder and CEO Max Russell-Harris is building around that lost context. Wrapped works inside the electronic health record, helping care teams assemble product recommendations they review and send to patients. The patient gets a ready-to-buy selection from retailers, with FSA/HSA funds available for eligible purchases.

The practice keeps its place in the decision without taking on inventory, payments or shipping, according to Wrapped's product materials. That is an important commercial choice: ask a provider to make the guidance easier to act on while keeping the retail operation elsewhere.

The investors have a clear view of that handoff. Lightbank partner Eric Ong describes the opportunity around clinician workflows. Russell Glass, Arteria Capital's founder and general partner and former Headspace CEO, points to convenience as a driver of whether people follow through.

Wrapped says it is live in clinics and available through the athenahealth Marketplace to more than 180,000 clinicians on athenaOne. That figure describes potential distribution, not a disclosed count of Wrapped customers. Getting inside the software a practice already uses gives the company somewhere useful to start the adoption conversation.

The $2.1M will support product and engineering development, additional EHR integrations, and partnerships with provider organizations, health systems and digital health platforms. Each integration puts another care team's routine within reach, along with its expectations about what a recommendation should carry into the patient's day.

For the family at home, that last part is personal. The clinician has chosen something for a reason, and the shopping experience still has to preserve enough of that reason to help someone get the right product while caring for somebody they love.

Pathfinder

Amaan Ahmad and Sidharth Jain ran a microschool together in London before building a financial product for Arizona parents. That journey puts them close to a peculiar education problem: choosing the right resource for a child can be easier than getting the purchase through the funding system.

Their company, Pathfinder, has raised a $12M Series A led by Portage (Portage Ventures), with Reach Capital, Gradient and Stand Together Ventures Lab participating. The November 2 announcement brings total funding to $16M, following the $4M Seed co-led by Reach and Gradient.

Amaan Ahmad, CEO, and Sidharth Jain, CTO, are engineers who chose to spend time around families doing the work. Reach describes a product shaped in living rooms, churches, co-ops and coffee shops. Parents were already assembling an education. They needed the financial administration to keep up.

Pathfinder combines an education charge card with an app that helps check eligible purchases, prepare documentation and submit reimbursements to ClassWallet. Column, N.A. issues the Visa card. State-approved reimbursements pay down balances, giving families a way to make purchases while the administrative process runs.

There is a consequential relationship inside that arrangement. A parent is trusting a private company's guidance while spending money governed by a state program. A useful answer has to carry through to the receipt, the reimbursement request and the family managing the balance.

Pathfinder puts some of its own money behind that guidance. Its published guarantee covers qualifying purchases its tool gets wrong, subject to terms and caps of $2,500 per item and $10,000 per cardholder per year. The card is a charge card, with the full balance due each billing cycle. Those details belong beside the convenience story.

The company reports more than 8,000 Arizona ESA accounts since launching in March. That is company-reported adoption, with further expansion still ahead. New capital will support product development and a plan to reach at least six states in 2027 while pursuing money transmitter and servicing licenses.

For the investors, the opportunity extends into the repeated decisions families make about educational spending. Earning that position takes accurate guidance and a process parents can follow after checkout. Amaan Ahmad and Sidharth Jain are carrying their microschool experience into a much larger collection of households, each with a child to teach and a purchase to get approved.

Northwood Space

A satellite operator booking a contact through Northwood's API is also booking the work of a factory in Torrance. The reservation looks like software. Behind it sit antennas, permitted sites and people responsible for keeping a spacecraft connected.

Bridgit Mendler, co-founder and CEO, built Northwood Space with CTO Griffin Cleverly and Head of Software Shaurya Luthra around that whole assignment. Owning the ground network means taking responsibility for the less glamorous work a customer would otherwise have to assemble across suppliers.

Northwood has closed a $375M Series C co-led by Andreessen Horowitz and DFJ Growth. Washington Harbour Partners, Alpine Space Ventures, Stepstone and In-Q-Tel are among the returning investors. A separate HSBC Innovation Banking credit facility supports working capital and expansion; its undisclosed size stays separate from the equity raise.

The company makes Portal phased arrays for command and control and Prism parabolic antennas for high-throughput data. Its software coordinates the hardware and lets operators reserve, update or cancel satellite passes. The customer gets capacity on a shared network, with Northwood carrying the work of building and operating the ground underneath it.

That puts a manufacturing decision inside a service business. At the new 185,000-square-foot headquarters and factory, Northwood is bringing more production and testing in-house. A circuit board delayed at a supplier can become a deployment delay, then a customer waiting for a link. Control over the build gives the team more of that calendar to manage.

Northwood reports six multi-orbit-capable sites across two continents and more than 140 employees. It plans to add 400 jobs over three years. Its earlier $100M Series B followed a $30M Series A, while a $49.8M Space Force award put the network close to demanding government missions. The award is a contract, with delivery work attached.

Shared infrastructure asks customers to trust capacity they do not own. It also asks Northwood to put equipment in the right places before every future reservation is known. Geography, mission requirements and available contact windows will shape how productively that investment gets used.

This round backs the factory, the network and the growing team expected to make those pieces behave like one service. As Northwood adds sites, the satellite operator will keep returning to the same simple request: reserve the next pass.

nPhase

The human approval step survives inside nPhase's agentic AI pitch. Scott Climes and Jaya Prakash Rao are expanding the software's assignment around it, from organizing a study's data to preparing the outputs a pharmaceutical sponsor will submit.

That is a larger assignment than selling a place to store study records. Scott Climes, CEO, and Jaya Prakash Rao, CTO, are bringing data management and statistical computing into the same commercial relationship. The sponsor gets a supplier asking to own more of the journey, including the awkward handoffs where somebody usually has to reconstruct what somebody else meant.

On November 2, nPhase announced a $30M Series D led by Viking Global Investors. The company calls it its first equity raise since 2017, after nearly a decade building without additional equity. That interval gives the financing a different texture: capital is arriving behind an operating history and a broader product assignment.

The history matters. REDCap Cloud grew out of a Vanderbilt licensing collaboration. By 2019, nPhase was adding analytics and warehousing across research studies. September's nPhase.ai launch joined its Unified Data Management System, or UDMS, with its Statistical Computing Environment, or SCE, extending the work toward submission-ready outputs. Connecting collection to analysis asks teams with different jobs to agree on the same evidence and its history.

The company's SCE demonstration makes that division of responsibility explicit. AI proposes transformations, humans approve them, and deterministic engines execute them. The people responsible for the submission keep authority while the software takes on more preparation. A sponsor has to be able to explain the result after the impressive demo has ended.

nPhase says its platform can compress submission preparation that typically takes four or more months into weeks. That is a company performance claim. For a buyer, the commercial question also reaches into the review process: can programmers inspect the work, reproduce it, and carry the record forward when the study changes?

Viking is backing expansion of the enterprise business and agentic AI capabilities across that clinical-data lifecycle. As more preparation moves into one platform, Climes and Rao are taking on a continuing relationship with the sponsor's reviewers, who still have to put their names behind the package leaving the system for regulatory review.

TailorMed / GrowthCurve Capital

Srulik Dvorsky and Adam Siton could have kept TailorMed focused on finding help with the bill. They have been expanding the job around it, following the patient through access, affordability and the work of staying on medication.

That choice now has a new majority investor. GrowthCurve Capital's November 2 investment backs a company whose usefulness depends on connecting organizations that see different parts of the same treatment journey. The companies have not disclosed financial terms.

Srulik Dvorsky, TailorMed's co-founder and CEO, describes a business that began with affordability and kept moving across the medication lifecycle. Adam Siton, co-founder and CTO, has a demanding technology assignment inside that ambition: make the connections useful to the people who have to complete the work.

A financial assistance program can exist, have funding and still miss the patient who qualifies. Somebody has to identify that person, match the requirements, get the application through and stay close enough to notice when support needs renewing. A full directory of programs leaves plenty of work on the care team's desk.

TailorMed Core puts financial navigation into care-team workflows. Connect gives patients a way to find support and complete enrollment themselves. The wider platform adds expert services, manufacturer solutions and partner offerings, extending the relationship beyond the first application.

TailorMed says its network spans 950+ hospitals, 4,700+ clinics and 3,100+ pharmacies. It reports more than $7.4B in assistance secured since 2020. That figure describes assistance secured through the network; it is not the company's revenue or the size of GrowthCurve's investment.

GrowthCurve's Vignesh Aier, head of healthcare investing, and director Robert Thwaits are backing a network their firm can help develop through AI, product and engineering. Established relationships give additional capabilities somewhere to go.

For a hospital or pharmacy buying this software, the appeal is practical. Staff need to know who is eligible, where an application stands and which task still belongs to a person. Patients need the resulting support to arrive while the prescription is still waiting to be filled.

Expanding across the medication journey gives TailorMed more places to help and more handoffs to keep intact. Srulik Dvorsky and Adam Siton are carrying that responsibility into a majority-owned chapter, with care teams still working through the next enrollment, renewal and delayed start.

Specific Biologics

Brent Stead went back to his old PhD examiner in 2017 with a commercialization idea. David Edgell had the gene-editing research; Brent Stead wanted to carry it into medicines, with all the years of company-building that would require.

That relationship sits behind Specific Biologics, the Toronto biotech that has closed a US$35M Series A led by Hatteras Venture Partners. Lumira Ventures and adMare BioInnovations are returning, joined by Genesys Capital, BDC's Life Sciences Venture Fund, DeepWork Capital and Seido Capital.

The scientific assignment is unusually specific. In C9ORF72-associated ALS, an expanded stretch of repeated DNA is part of the disease's genetic machinery. Specific's Dualase platform is designed to shorten that expansion while preserving the normal sequences the cell needs. Removing the wrong thing would rather defeat the purpose.

Co-founder and CEO Brent Stead and CSO Brad Geddes are carrying that approach toward clinical development. The financing is intended to take the lead candidate through initial clinical biomarker proof of concept in the first cohort of an adaptive Phase I/II trial. That gives the round a specific early milestone, with human safety and benefit still to establish.

Delivery belongs inside the investment conversation. Specific says its compact editor and guides fit into a single adeno-associated virus, or AAV, for delivery to the central nervous system. Packaging changes the manufacturing and dosing assignment long before anyone can assess what a treatment does for a patient.

Specific reports repeat reduction in patient-derived cells and animal models, with no detectable off-target effects in those preclinical tests. Those findings explain the financing thesis. They remain evidence from models, and the announcement gives no trial-start or patient-enrollment date.

Lumira and adMare backed the first institutional financing in 2021 and invested again in 2023. Their return alongside a new lead carries an academic-origin platform toward a different kind of work, where clinicians and patients enter a story previously told through laboratory results.

For a biotech operator, that continuity matters. The research relationship has survived the trip from a university office to successive financing rounds. Brent Stead's next conversations will have to carry the precision promised by Dualase into the choices surrounding that first clinical cohort. A biomarker readout would then inform how the program moves beyond that cohort.

&Truth

A product owner signing off on an AI agent's spending limit has joined the verification business. The rule has to survive its trip from a business decision into the software that handles the money.

&Truth is building for that trip. Founded in early 2026 by Chris Hsu, Clark Barrett and Ilya Sergey, the public benefit corporation emerged from stealth on October 29 with $70M in seed capital. Its investors were described as AI safety and venture capital leaders, without names or a disclosed valuation.

Chris Hsu and Clark Barrett are co-CEOs; Ilya Sergey is chief science officer. Clark Barrett brings Stanford automated-reasoning research, while Ilya Sergey founded the VERSE lab at the National University of Singapore. The company operates from Palo Alto and Singapore, with further Singapore expansion planned across research, engineering, product and go-to-market.

The interesting commercial assignment sits inside what &Truth calls a specification factory: translating human intent and business requirements into precise formal specifications. A developer can produce code quickly. Deciding exactly which actions that code may take still requires somebody who understands the business, the permissions and the damage an ambiguous instruction could cause.

&Truth is developing a verification machine that checks software against those specifications and produces mathematical proofs others can inspect. Its foundations include Lean and CSLib, with foundational libraries intended for open-source release. Cross-language verification is the ambition; the announcement supplies no complete coverage matrix or independently validated performance benchmark.

That makes the specification a serious product decision. A formally correct spending rule can still leave the buyer with the wrong spending policy. The proof carries the requirement it was given, under stated assumptions, so choosing and maintaining that requirement belongs in the same conversation as building the software.

For a security leader or engineering buyer, independent checking changes the proposed relationship with the vendor. The buyer would have evidence to inspect beyond a reassuring dashboard or another assurance that the system was tested. Somebody must also keep that evidence aligned with the version of the software actually being deployed.

The $70M backs an attempt to make this work repeatable across legacy and AI-written code. As &Truth develops that vocabulary, the product owner still has a signature to put on the rule that tells an agent when to stop spending.

Applied Compute

Clay once asked users to choose the AI model for each Claygent task. Now a trained router can make that choice query by query, putting a surprisingly consequential purchasing decision inside the product itself.

A contact lookup and a difficult company investigation can arrive through the same interface. Paying for the strongest model on both is a comfortable habit for the vendor collecting the bill. Clay and Applied Compute built a router around the harder question: how much capability does this particular request need?

Applied Compute says that work cut Claygent inference costs by up to 47% while maintaining quality. The case study describes guardrails for individual query categories, designed to keep aggregate savings from borrowing accuracy from the awkward jobs. That detail belongs in the buying conversation.

CEO Yash Patil, CTO Rhythm Garg and Chief Architect Linden Li have built the company around these specific decisions. The three co-founders previously worked at OpenAI on Codex, reasoning and reinforcement-learning infrastructure. Their current assignment reaches from training a customer's model to running it and improving it once real work starts coming through.

On November 2, Applied Compute announced $350M in new funding at a $3.25B valuation, led by Elad Gil. Kleiner Perkins, Lux Capital, Hanabi Capital, NVIDIA, Blackstone, Shopify Ventures and HubSpot Ventures are among the participants. The money is intended for compute, hiring and enterprise deployment.

The company reports a $100M annualized revenue run rate in under 18 months. That is a current revenue pace, not a completed year's revenue. Its customers include Microsoft, Harvey, DoorDash, Mercor and Cognition, alongside Clay.

AC2 brings training, serving and continued improvement onto one platform. Customers can bring their own training harnesses and graders into AC2. For a buyer, that means the evaluation criteria become part of a recurring operating relationship: somebody has to decide what good work looks like, inspect failures and keep those definitions useful as the product changes.

Applied Compute says 40% of customers now use it for both training and inference. There is a commercial ambition inside that overlap: help build the model, then earn a continuing place in the work it performs.

For Clay, the conversation keeps returning to the next query in the list, and whether its router knows when a cheaper answer is good enough.